Download 2016 Conference of Asia Pacific Association for Medical Informatics

C o n nec ti ng t h e Do t s fo r Better Hea lth thro u g h Info rma t ics
2016 Conference of
Asia Pacific Association for
Medical Informatics
November 2 (Wed) – 5 (Sat), 2016
The K-Hotel, Seoul, Korea
Sponsored by
1
Contents
Welcome Message
Organizing Committee
Advisory Committee
Scientific Program Committee
Keynote
Plenary
Program at a Glance
Oral Session
Government Project Session
Poster Session
Social Program
Floor Plan
Exhibition
Venue Information
3
5
5
6
7
8
9
11
15
16
18
19
20
21
Oral Session
ㆍsession1
Telemedicine for Global Health Network
22
ㆍsession2
Big Data Informatics
28
ㆍsession3
Health Data Exchange
33
ㆍsession4
Nursing Informatices
40
ㆍsession5
Personal Health Record (PHR) and Hospital System
45
ㆍsession6
Clinical Decision Support System
50
ㆍsession7
Electronic Health Record (EHR)
57
ㆍsession8
mHealth for Connected Patient Care
63
ㆍsession9
(APAMI Section) Public Health & National Health Strategy
68
ㆍsession10 Mobile Health Application
74
ㆍsession11 Artificial Intelligence in Medical Informatics
80
Poster Session
ㆍP1
86
ㆍP2
106
2
2016 Conference of
Asia Pacific Association for Medical Informatics
Welcome Message
I am Minister chung chin-youb of the Ministry of Health and Welfare of the Republic of Korea.
I would like to congratulate the successful hosting of 2016 conference of the Asia Pacific Association for
Medical Informatics (APAMI).
APAMI initiated this conference in 1994 to together explore a direction to develop medical informatics in Asia
Pacific. Ever since, it has taken place every year. I am all the more pleased to have this historic conference in
Korea this year. And I extend warm welcome to all of you here today.
The Korean Society of Medical Informatics (KOSMI) is the organizer of this conference. Since 1987, this
Society has contributed to produce quality healthcare services by converging tech and intelligence to meet
the demands of the information society. And it has enhanced public health standards.
In this regard, my special thanks go to Organizing Chair Kyung-Hee Cho and Ju-Han Kim, President RongMin Baek and chairman Jinwook Choi of the KOSMI, and other staff members for putting utmost efforts in
organizing this event.
I heard that this conference consists of vibrant discussions and presentations on various themes, including
telemedicine, big data, information exchange, standard, and mobile health, under the topic of “Connecting the
Dots for Better Health through Informatics”. I expect this conference will be an opportunity to share the future
direction of health informatics and mutual growth and cooperation.
Health informatics is a field of science that utilizes data on health maintenance, health care, and patient
medical records. This area is gaining more and more importance because the health sector is not immune
to the global IT advancement. Population aging is boosting public interest in healthy longevity. Healthcare
paradigm is shifting from treatment to prevention.
Under this circumstance, IT health policy tops the list of core national tasks worldwide. Korea is expanding
ICT-based medical services such as telemedicine and health information system,
and making R&D investment to promote precision medicine using big data, genomic
information, and AI.
Korea retains abundant medical big data to be used in disease research and new
industry creation. To make the most use of it, Korea is gradually opening national
health insurance data to the public.
I hope this conference will serve as a venue for active discussions to bring health
informatics to the next level and share open-source health information and
technology of the Asia-Pacific countries.
Again, I appreciate all the distinguished participants and APAMI members. Wish you
happiness and wellbeing.
Thank you.
3
Chung Chinyoub
Minister of Health and
Welfare
Welcome Message
Connecting the Dots for Better Health through Informatics, this conference’s theme, aims to encourage
researchers, surgeons, practitioners, young scientists, healthcare workers and suppliers in the world to
cooperate and share knowledge and experiences each other about how to use medical technology to
improve medical information of the people.
Distinguished guests of the government, business, IT and academic communities will be invited to our
conference to present and share their knowledge and experiences. We estimate more than 400 global
participants to come to our conference and discussions for the development and application of new
technologies in medical and health informatics will be progressed by our prominent speakers.
We will try to provide not only outstanding scientific programs and social events for you and also
excellent opportunities to make new relationship with people related to medical and health informatics.
As our efforts, we organize a tour program including visiting Korean representative smart hospitals and
experience of Korean traditional culture to make your enjoyable and memorable moment during staying in
Korea.
We are looking forward to seeing you here.
Dr. Kyung-Hee Cho
Dr. Ju-Han Kim
Organizing Chair
Organizing Chair
4
2016 Conference of
Asia Pacific Association for Medical Informatics
Organizing Committee
Category
Chair
Secretary General
Scientific Program
Publication
Name
Kyung-Hee Cho
Affiliation
National Health Insurance Service Ilsan Hospital
Ju-Han Kim
Seoul National University College of Medicine
Hasuk Bae
Ewha Womans University
Jae-Ho Lee
Asan Medical Center
In-Young Choi
Hyoun-Joong Kong
Seol-Kyung Baek
The Catholic University of Korea
Chungnam National University
Ajou University Hospital
Finance
Younghee Lee
Halla University
PR & Sponsorship
Seung-Su Kim
Yonsei University Health System
Social Program
Young Ah Kim
Yonsei University Health System
Registration
Soo-Yong Shin
Asan Medical Center
Il Kon Kim
Committee
Kyungpook National University
Hyeoun-Ae Park
Seoul National University
Jinwook Choi
Seoul National University
Hun Sung Kim
The Catholic University of Korea
Advisory Committee
Category
Advisory
Committee
Name
Myeung-Ki Kim
Seoul National University
Suk-Wha Kim
Seoul National University
Hyeoun-Ae Park
Seoul National University
Rong-Min Baek
Seoul National University
Kun-Ho Yoon
The Catholic University of Korea
Kap-No Lee
Seegene Medical Foundation
Young-Sung Lee
Chungbuk National University
Byung-Chul Chang
Hune Cho
5
Affiliation
Yonsei University
Kyungpook National University
Scientific Program Committee
Name
CP Wong
Sushil Kumar Meher
Wu-Chen Su
Vajira H. W. Dissanayake
Hong Kong Society of Medical Informatics
India Institute of Medical Sciences, New Delhi
University of Kentucky
Health Informatics Society of Sri Lanka (HISSL)
Rakesh Kumar Das
B. P. Koirala Institute of Health Sciences, Dharan, Nepal
Adam Chee
Association for Medical and Bio-Informatics, Singapore
(AMBIS)
Michio Kimura
Zhi Yang
Scientific Program
Committee
Affiliation
Hamamatsu University
Capital Medical University, Professor
Department of Biomedical Instrumentation
School of Biomedical Engineering
Hun Sung Kim
The Catholic University of Korea, College of Medicine
Dukyong Yoon
Department of Biomedical Informatics, Ajou University
School of Medicine, Suwon, Korea
Sun-Mi Lee
The Catholic University of Korea, College of Nursing
Hyung Jin Choi
Seoul National University
Dong Kyun Park
Gachon University Gil Medical Center
Chi-Ming Chu
National Defense Medical Center
Won Chul Cha
Department of Emergency Medicine, Samsung Medical
Center
Shashi Bhushan Gogia
Joo-Hee Park
Hwa Jeong Seo
In-Sook Cho
Hong-Gee Kim
Seung-Hyeon Jung
Yul-Ha Min
Society for Administration of Telemedicine and Healthcare
Informatics (SATHI)
Sahmyook Health University
Gachon University
Inha University
Seoul National University
National Cancer Center
Gachon University
6
2016 Conference of
Asia Pacific Association for Medical Informatics
Keynote
The Role of Informatics in Digital Health: Integration of individual Genotypes,
Phenotypes and Expotypes
Fernando MartinSanchez PhD,
FACMI, FACHI
Prof., Division of Health
Informatics, Department of
Healthcare Policy and
Research, Weill Cornell
Medicine.
New York. USA
The term ‘digital medicine’ is increasingly being used synonymously with areas such
as mobile health, digital health, health IT, or health 2.0. Whilst digital medicine
encompasses numerous different concepts, i.e. quantified self, sensors, apps, telehealth, games, electronic health records, interoperability, etc., we are yet to see
how Biomedical Informatics is included in this new landscape. If we consider digital
medicine to be at the intersection between Biomedical Informatics, Participatory
Medicine and Precision Medicine, we can identify new critical roles for our research
including self-quantification and the use of social media, as well as new approaches
to process, integrate and analyse data originating from the Human Genome, Phenome
and Exposome. It could be beneficial to consider a new way of structuring our skills,
methods, expertise and educational programs such that biomedical Informatics is
placed as the discipline that can deal with all these types of data, as well as with their
interplay. Hence, our relevance should clearly be maintained in areas of application like
the aforementioned.
A Phone is Worth a Thousand Miles
Yu-Chuan (Jack) Li,
MD, PhD
FACMI Professor and Dean,
College of Medical Science
and Technology, Taipei Medical
University, Taiwan
Many developing countries suffer a scarcity of trained clinicians, who are usually
concentrated in urban centres, leaving large rural populations essentially underserved.
The adoption of technology could offer new opportunities for patient benefit in term of
costs, better care and in turn, better outcomes. Mobile health technology is promising
to ameliorate the widening healthcare supply and demand gap through which the
reach of consultants can be digitally extended, allowing them to virtually cover larger
or remote regions. Smartphone applications have been shown to be a simple, feasible
and reliable method for performing store-and-forward teledermatology consultations in
rural unpopulated areas. This talk would focus on several m-health applications and their
outcomes.
Open Source Strategy for Health Information Technology
Seong K. Mun, PhD
Prof., Virginia Tech and
OSEHRA, Arlington, VA, USA
The essence of open source software operation is unimpeded collaboration. Successful
open source ecosystem consists of software, governance and community. These
three components have to work harmoniously to support various business objectives
of participants. The largest open source organization in health IT is the Open Source
Electronic Health Record Alliance (OSEHRA) established by the US government as
an independent non-profit corporation. It is the home the most comprehensive open
source EHR, known as VistA which are used more than 2,000 facilities around the
world. Recently a major new modern development and testing environment known as
enterprise health management platform (eHMP) has been released through OSEHRA o
the global community to promote rapid innovation in VistA through collaboration.
Adoption of open source practice is increasing rapidly throughout the software industry.
Though open source is relatively new concept in health IT, time has come to look at open
source strategy to make health IT readily available to all communities and countries and
a way of controlling users’ destiny.
Many computer codes are developed especially in research labs but most of them
become useless in long run. Open source strategy offers sustained development and use
of the codes beyond research projects. This session will focus on open best practices,
useful standards, and business models that can take advantage of open source concept.
7
Plenary
Medical Informatics Year in Review
Hyeoun-Ae Park,
RN, PhD
Prof., Seoul National University
College of Nursing, Seoul,
Korea
The field of biomedical informatics is advancing drastically, especially in research and
practice. In order to highlight the accomplishments of the research work, elicit research
trends and identify research patterns of biomedical informatics at a macro level, a
7-person team conducted an extensive review of the literature on clinical and consumer
health informatics published from October 2015 to September 2016. Journals reviewed
include International Journal of Medical Informatics, Journal of American Medical
Informatics Association, Journal of Biomedical Informatics, and Journal of Medical Internet
Research. Keywords used in the search include decision support systems, EHR, electronic
health records, m-health, consumer informatics, public health informatics, precision
medicine, meaningful use, patient reported outcomes, clinical informatics, medical
informatics, big data initiative, health information exchange, telemedicine, evaluation of
EHR, consumer health informatics, precision medicine, and health informatics.
Research topics will be organized under 3 themes: the electronic health records (EHR),
consumer health informatics, and the learning health system. Researches will also be
grouped under 3 different research methodologies: quantitative, qualitative, and mixed
methods.
Key findings include the following: There are significant advances in EHR implementation,
however interoperability of EHR data is still an issue to fully realize the benefits of the
EHR. Decision support systems are being implemented to improve clinical practice,
however evidence of the impact on patient outcomes is still lacking. Mobile applications
are being introduced to improve consumer engagement, however evidence of the impact
on patient outcomes remains elusive. Big data are receiving increasing attention in
healthcare, especially translational aspect, however data access and privacy remain
issues. Quantitative methods are being used the most frequently, however we are
beginning to see more use of mixed methods in biomedical informatics research.
Realizing Precision Medicine through Translational Bioinformatics
Ju-Han Kim,
MD, PhD, MS
Prof., and Founding Chair, Div.
of Biomedical Informatics,
Seoul National University
College of Medicine, Seoul,
Korea
A flood of multi-modal high throughput clinical genomic data and personal health
records means that many of the challenges in biomedical research and healthcare
are now challenges in integrative and computational sciences for their bidirectional
translations. Our ability to ‘connecting the dots’ in the wealth biomedical big data will
bring us the ‘big picture’ in a mass of genes, drugs, diseases, and diagnostic, therapeutic
and prognostic markers. Precision medicine attempts to determine individual solutions
based on the genomic and clinical profiles of each individual, providing opportunity to
incorporate individual molecular data into patient care. While a plethora of genomic
signatures have successfully demonstrated their predictive power, they are merely
statistically-significant differences between dichotomized phenotypes that are in fact
severely heterogeneous. Despite many translational barriers, connecting the molecular
world to the clinical world and vice versa will undoubtedly benefit human health in the
near future.
8
2016 Conference of
Asia Pacific Association for Medical Informatics
Program at a Glance
November 3 (Thu), 2016
Time
Program
08:00-
Registration
(Geomungo Hall Lobby)
09:00 - 9:30
Opening Message: Organizing Chair of APAMI 2016
Congratulatory Remarks: President of APAMI (Shasi Bhushan Gogia)
President of KOSMI (Rong-Min Baek)
President of IMIA (Hyeoun-Ae Park)
[Moderator]
Jae Ho Lee, Korea
(Geomungo A, B, C Hall)
09:30 - 10:30
Keynote: The Role of Informatics in Digital Health: Integration of individual Genotypes,
Phenotypes and Expotypes
(Fernando Martin-Sanchez, Weill Cornell Medicine, USA)
[Moderator]
Hyeoun-Ae Park, Korea
(Geomungo A, B, C Hall)
10:30 - 10:50
Coffee Break (Geomungo Hall Lobby)
Session 1
Telemedicine for Global Health
Network
10:50 - 12:30
[Moderator]
Suk Wha Kim, Korea
(Geomungo A)
12:30-12:50
Session 2
Big Data Informatics
Session 3
Health Data Exchange
[Moderator]
Oommen John, India
[Moderator]
Michio Kimura, Japan
Il Kon Kim, Korea
(Geomungo B)
(Geomungo C)
Invited Session: Changing Paradigm of Critical Care ICT: eICU
(Sueyoung Yoon, Philips)
[Moderator]
Byung Chul Chang, Korea
12:50-14:00
Luncheon (Geomungo A, B, C Hall)
Plenary: Medical Informatics Year in Review
(Hyeoun-Ae Park, Seoul National University, Korea)
14:00-15:00
15:00 - 16:00
16:00 - 17:30
17:45 - 18:45
9
[Moderator]
Shasi Bhushan Gogia, India / Rong-Min Baek, Korea
(Geomungo A, B, C Hall)
Coffee Break & Poster Presentation (Poster session 1)
(Geomungo Hall Lobby / Geomungo C Hall)
Session 4
Nursing Informatics
Session 5
Personal Health Record (PHR)
and Hospital System
Session 6
Clinical Decision Support
System
[Moderator]
jeongeun Kim
[Moderator]
Adam Chee, Singapore
Soo Yong Shin, Korea
[Moderator]
Rae Woong Park, Korea
(Geomungo A)
(Geomungo B)
(Geomungo C)
Transfer from The-K Hotel to Korean House
Program at a Glance
November 4 (Fri), 2016
Time
Program
08:00-
Registration
(Geomungo Hall Lobby)
Keynote: A Phone is Worth a Thousand Miles
(Yu-Chuan (Jack) Li, Taipei Medical University, Taiwan)
09:00-09:50
[Moderator]
Jin Wook Choi, Korea
(Geomungo A, B, C Hall)
Keynote: Open Source Strategy for Health IT
(Seong K. Mun, Virginia Tech and OSEHRA, USA)
09:50-10:30
[Moderator]
In Young Choi, Korea
(Geomungo A, B, C Hall)
10:30-10:50
Coffee Break (Geomungo Hall Lobby)
10:50-12:30
Session 7
Electronic Health
Record (EHR)
Session 8
mHealth for Connected
Patient Care
[Moderator]
Seong K. Mun, USA
Kap No Lee, Korea
[Moderator]
Zhi Yang, China
Hune Cho, Korea
(Geomungo A)
(Geomungo B)
12:30-14:00
Government
Project 1
IoA3 for Health:
Internet of Avatars,
Agents, and Apps for
Health
11:00-12:00
APAMI
General Assembly
(Language: Korean)
(Geomungo C)
(Haegeum B)
Luncheon (Gayageum A, B Hall)
Plenary: Translational Bioinformatics for Precision Medicine
(Ju-Han Kim, Seoul National University College of Medicine, Korea)
14:00-15:00
15:00-16:00
[Moderator]
Kyung Hee Cho, Korea
(Geomungo A, B, C Hall)
Coffee Break & Poster Presentation (Poster session 2)
(Geomungo Hall Lobby / Geomungo C Hall)
Session 10
Mobile Health
Application
Session 11
Artificial Intelligence in
Medical Informatics
[Moderator]
C P Wong, Hong Kong
Jae Ho Lee, Korea
[Moderator]
Rakesh Kumar Das,
Nepal
Kun Ho Yoon, Korea
[Moderator]
Naoki Nishimoto, Japan
Dong Kyun Park, Korea
(Geomungo A)
(Geomungo B)
(Geomungo C)
Session 9
(APAMI Section)
Public Health & National
Health Strategy
16:00-17:30
17:30-18:00
Government
Project 2
Research Platform
as a Public &
Government Service
using Health BigData
(Language: Korean)
(Haegeum B)
Best Paper & Poster Award
APAMI 2018 Promotion (President of APAMI)
Medinfo 2017 Promotion
Closing Remarks
[Moderator]
In Young Choi, Korea
10
2016 Conference of
Asia Pacific Association for Medical Informatics
Oral Session
Category
November 3 (Thu), 2016
Time
Abstract Title
10:50-12:30
Moderator: Suk Wha Kim, Korea
10:50-11:15
1-1
Using eHealth for Addressing Lymphoedema
in India
Arun Gekha Gogia,
India
Session 1
Telemedicine for 11:15-11:40
Global Health
Network
1-2
Assessing the implementation of the National
e-Health Strategic Planning to support
telemedicine projects in Indonesia
Sumarsono
Surahyo,
Indonesia
11:40-12:05
1-3
Informatics beyond boundaries: An experience
of technology enabled re-mote support in Natural
Disasters (Nepal Earthquake 2015 Relief Portal)
Oommen John, India
12:05-12:30
1-4
Telemedicine setup and consultation process
in a remote health post in Eastern Nepal
- a cross sectional study
Rakesh Kumar Das,
Nepal
(Geomungo A)
10:50-12:30
Session 2
Big Data
Informatics
(Geomungo B)
Session 3
Health Data
Exchange
(Geomungo C)
Moderator: Oommen John, India
10:50-11:10
Invited
Session
Big Data for Clinical Informatics with Oracle Solution
Jeyaseelan Jeyaraj,
ORACLE
11:10-11:30
2-1
Data Validation Study for Japanese Sentinel Project
Rieko Izukura, Japan
11:30-11:50
2-2
A Comprehensive Map of Polygenic Susceptibility
Associations Among Diseases and Traits Based
Jung Hun Ohn, Korea
on Large-scale Individual Genotype Data
11:50-12:10
2-3
FCM-EM: Discovering DNA Motifs by Machine
Learning Techniques
12:10-12:30
2-4
Development a minimum data set of the information
management system for thalassemia
10:50-12:30
11
Presenter
Ali Yousefian, Korea
Leila keikha, Iran
Moderator: Michio Kimura, Japan
Il Kon Kim, Korea
Experience of Korean Hospitals in Health
Information Exchange based on Standards
Jeung-Whun Kim, Korea
3-1
Implementation of standards-based, vendor
neutral archive-driven API platform and client
Jung Bin Kim, Korea
11:30-11:50
3-2
Unlocking Interoperability Puzzle with the Global
Language of Health - SNOMED CT
Michael Bainbridge,
Australia
11:50-12:10
3-3
Pilot study for a mapping from HL7/SS-MIX2 to
CDISC SDTM
Naoki Nishimoto,
Japan
12:10-12:30
3-4
Data Quality Management for Clinical Contents
Model
Yookyung Boo, Korea
10:50-11:10
Invited
Session
11:10-11:30
Category
Time
Abstract Title
16:00-17:30
Session 4
Nursing
Informatics
(Geomungo A)
4-1
Develop and validate an empowerment tool for
pregnant women
16:30-17:00
4-2
The development and acceptability evaluation of
the virtual simulated hospital training program
for nursing students
17:00-17:30
4-3
The Need Assessment of Long-Term Care
Information in Taiwan
Rita Rezaee, Iran
Ying Wu, China
Abdulla Ahmed
Elbakkoush, Taiwan
Moderator: Adam Chee, Singapore
Soo Yong Shin, Korea
PHR in NHIMC
Hee-Jin Kang, NHIS
Korea
5-1
Development of Personal Health Records
Architecture for Community Oriented Holistic
Care to be in Line with International Standards
Heng-Shuen Chen,
Taiwan
5-2
Development of a Mobile Application in
Conjunction with a Web Service for Personal
Health Records regarding Atopic Dermatitis
16:54-17:12
5-3
Dynamic relationship between emotion and
physical states: an observational study through
personal health record
17:12-17:30
5-4
From EHR to PHR: a Literature Review
16:00-16:18
Session 5
16:18-16:36
Personal Health
Record (PHR)
and Hospital
System
16:36-16:54
Invited
Session
(Geomungo B)
16:00-17:30
Session 6
Clinical Decision
Support System
Moderator: jeongeun Kim
16:00-16:30
16:00-17:30
Presenter
Bo-Hyeong Jang,
Korea
Ye Seul Lee, Korea
Ching Chu Liao,
Taiwan
Moderator: Rae Woong Park, Korea
16:00-16:18
6-1
Classification of clinical prediction rules by
design and usage
16:18-16:36
6-2
Association of lipid profile & HbA1c among
T2DM patients: A cross-sectional study in
Bangladesh
Md.Mohaimenul
Islam, Taiwan
16:36-16:54
6-3
The potential use of computer-assisted image
processing for detecting Acute Lymphoblastic
Leukemia
Lutfan Lazuardi,
Indonesia
16:54-17:12
6-4
EMRs- Towards a better User Interphase (UI)
Shashi Bhushan
Gogia, India
17:12-17:30
6-5
Impact of a QQ-based Virtual Health Paradise on
reducing cardiovascular disease risk factors
among healthy Chinese adults
Ying Wu, China
(Geomungo C)
Georgie Kennedy,
Australia
12
2016 Conference of
Asia Pacific Association for Medical Informatics
Oral Session
Category
November 4 (Fri), 2016
Time
Abstract Title
10:50-12:30
Moderator: Seong K. Mun, USA
Kap No Lee, Korea
Strategies for the Standardization of Electronic
Medical Record (EMR) Systems in Korea
Youngtaek Park,
HIRA Korea
10:50-11:10
Session 7
Electronic
Health
Record (EHR)
Invited
Session
11:10-11:30
7-1
The Overview of Free/Libre and Open Source
Shinji
Software in Medical Domain, from Asia and the World Kobayashi, Japan
(Geomungo A)
11:30-11:50
7-2
Kamal Wasantha
Strengthen cancer surveillance in Sri Lanka by
Kumara
implementing Cancer Registry Informatics to Enhance
Seneviratne,
Cancer Registry Data Accuracy, Completeness, and
SriLanka
Timeliness
11:50-12:10
7-3
Determining Morbidity Pattern First Step Deal with
Big Data via Electronic Health Record of Primary
Healthcare in Indonesia
12:10-12:30
7-4
An Viewing System for Health Check-up Result
Report using FHIR, CDA IG, and MHD
10:50-12:30
Session 8
mHealth for
Connected
Patient Care
Dina Nur
Anggraini
Ningrum, Taiwan
Mingyu Kim,
Korea
Moderator: Zhi Yang, China
Hune Cho, Korea
10:50-11:15
Invited
Session
Digital Healthcare Transformation with Mobility and IoT
Hwa Hyeon Yu,
HP Korea
11:10-11:30
8-1
Improving Patients’ Medication Adherence and Lipid Level
by Customized Short Message: a Six-Month Clinical Trial
Yisi Liu, China
11:30-11:50
8-2
Impact on Effectiveness of Chronic Disease
Management in a Rural Hospital of middle Taiwan
by a Case Management Information System
Mei-Chih Chen,
Taiwan
11:50-12:10
8-3
Remote Monitoring in Home Peritoneal Dialysis:
Seamless User-centered Proactive Provision of
Risk-stratified Treatment in Peritoneal Dialysis
(SUPPORT-PD)
Oommen John,
India
12:10-12:30
8-4
Disruptive Innovation in mHealth: A case of Nutrition
Monitoring from a LMIC
(Geomungo B)
13
Presenter
Pamod Madusanka
Amarakoon,
SriLanka
Category
Time
Abstract Title
16:00-17:30
History of Health IT in Japan
- Invitation to IMIA’s 50th anniversary -
9-2
Climate & Health Innovation for Substainable
Smart Cities
Adam Chee,
Singapore
9-3
The development and implementation of Stroke risk
prediction model in National Health Insurance’s
Personal Health record
Jae-Woo Lee,
Korea
16:54-17:12
9-4
Dialysis Outcomes in India: An online clinical
outcomes registry for dialysis outcomes, lessons
learned and opportunities for scale up
Oommen John,
India
17:12-17:30
9-5
State of Management Information System(MIS) in
Tertiary Care Hospitals in Pakistan
Suleman Atique,
Taiwan
Session 9
(APAMI Section) 16:18-16:36
APAMI and
National Health
Strategy
16:36-16:54
(Geomungo A)
16:00-17:30
(Geomungo B)
(Geomungo C)
Michio Kimura,
Japan
Moderator: Rakesh Kumar Das, Nepal
Kun Ho Yoon, Korea
16:00-16:20 10-1
Experience and Preference Factor Analysis for
pediatric obesity management mobile app.
16:20-16:40 10-2
Efficacy of Health Management in Rural Area
affecting by a Personal Health Records System with
Classification Algorithms
Hung-Pin Hou,
Taiwan
16:40-17:00 10-3
Analysis of a mobile electronic medical record usage
pattern: User interaction and evolution
Yura Lee, Korea
17:00-17:20 10-4
Utility of mobile apps in smoking cessation:
A literature review
Shwetambara
Vinayak Kekade,
Taiwan
16:00-17:30
Session 11
Artificial
Intelligence in
Medical
Informatics
Moderator: C P Wong, Hong Kong
Jae Ho Lee, Korea
9-1
16:00-16:18
Session 10
Mobile Health
Application
Presenter
Jisan Lee, Korea
Moderator: Naoki Nishimoto, Japan
Dong Kyun Park, Korea
16:00-16:20 11-1
Evaluating the effectiveness of distributional
hypothesis and name entity chunk tagging schemes:
a case of HMM for clinical named entity extraction
Wangjin Yi,
Korea
16:20-16:40 11-2
Coordinated Markov Modeling of Cancer Metastasis
from Multiple Primary Sites
Hyunggu Jung,
USA
16:40-17:00 11-3
Structuralization of Variance Text Records in Clinical
Pathway
Takanori
Yamashita,
Japan
17:00-17:20 11-4
Identify factors affecting successful ageing in elderly
diabetics in shiraz using data mining approach
Azam Aslani,
Iran
14
2016 Conference of
Asia Pacific Association for Medical Informatics
Government Project Session
Category
Government
Project
Time
10:50-12:30
November 4 (Fri), 2016
Abstract Title
Presenter
IoA3 for Health: Internet of Avatars, Agents,
and Apps for Health
Ju Han Kim, Korea
(Geomungo C)
Moderator: Young Sung Lee
Government
Project
Gov-1
The biosignal database construction for deep
learning based disease prediction
Dukyong Yoon,
Korea
Gov-2
Modeling of prognosis prediction and
management system for stroke using big data
Tea Jung Kim,
Korea
Development of Individualized Preventive
Management Service Model through
Implementation of Disease Risk Atlas and Risk
Prediction of Korean Common Chronic Diseases
Suekyung Park,
Korea
Gov-5
LOD typed global healthcare big data platform
development
Wan-Sup Cho,
Korea
Gov-6
Basic research on link application of healthcare
big data
Myung Jin Jeong,
Korea
15:00-17:30 Gov-3
(Haegeum B)
Gov-7
15
Discussion
Poster Session
NO
November 3 (Thu), 2016
Abstract Title
Presenter
Winnie Hoi Wing Leung,
Australia
1
Online Medical Appointment System
2
Comparison of tele-rehabilitation vs face-to-face vs usual care based
on community for low income elderly women in Korea
Jeongeun Kim, Korea
3
Hyoid Bone Tracking using Tracking-Learning-Detection in
Videofluoroscopy Swallowing Study
Dong Heon Lee, Korea
4
Characterizing hidden rules linking symptoms and acupoint selection
using an artificial neural network model
Wonmo Jung, Korea
5
Deep Belief Network based Classification Prediction Model of Pathology
Stage in Patients with Prostate Cancer
In Young Choi, Korea
6
Developing novel algorithms of approximate entity extraction for
laboratory test names in clinical documents based on q-gram
Kyung mo Kim, Korea
7
Development of new extravasation alert system by optical hemoglobin
sensor in drip infusion therapy
8
Improving of Medical Supply Purchasing Process by Using Barcode
Technology in Thai University Hospital
Laddawan Riponchai,
Thailand
9
AURORA: Analytic code managing and distributing Unit bridging Research
Organizations and Associated users
Junghyun Byun, Korea
10
Exploring the possibility of identifying adverse-drug-events from nursing
records
Eunjoo Jeon, Korea
11
A method of compressing whole slide images (WSI) with multi-focus plane
using blur estimation
Manwon Hwang, Korea
12
Suggestion of a Phased Classification Method of Requirements for
a Registry by the types of Cancer and the Definitions of Items using
Electronic Medical Records
Yeong Mi Jang, Korea
13
Development of Patient Portal for Exchange Personal Health Record
Ji hae Kwon, Korea
14
FHIR for IHE LAW Profile
Sunghwa Lee, Korea
15
Design of middleware solution for the vendor neutral archive system
Min Soo Park, Korea
16
Design and Implementation of SMART guide system based on
integrated data modeling
Young Hee Lim, Korea
17
Design and Development of a Medical Information Support System
Peixiang Yang, China
Hiroyuki Nozaka, Japan
16
2016 Conference of
Asia Pacific Association for Medical Informatics
Poster Session
17
November 4 (Fri), 2016
NO
Abstract Title
Presenter
1
Exploring factors affecting the time to fall from admission using EHR data
Hyesil Jung, Korea
2
The development of eHR domain content in HKSAR
Hung Hung TSUI,
Hong Kong
3
A Fast Microbial Detection Algorithm based on High-Throughput
Sequencing Data
4
Patient Engagement Function Usage Trends in a Hospital-tethered Mobile
Personal Health Records
Jae-Ho Lee, Korea
5
User requirement analysis for hyperlipidemia management application
Joo Yun Lee, Korea
6
Identification of Strategies to Promote Positive Mental Health among
the Adolescents using a Mobile App
Ji-Hyun Lee, Korea
7
A Design of Smartphone App for Children’s Weight Control with Their
Parents
Ah Jung Byun, Korea
8
Prescribing Error Reducing After the Implementation of Computerized
Physician Order Entry (CPOE) in Thai University Hospital
Ratchadaporn
Soontornpas, Thailand
9
Computerized Physician Order Entry(CPOE) Improved Financial Outcome
in Thai University Hospital
Cholatip Pongskul,
Thailand
Dongsheng Zhao
(Xiaolei Wang), China
10
An Investigation Report of Citizen's attitudes toward the handling of
Electronical Medical and Medicine Information.
11
A Research Study on the Design of Secure HIS that Based on
International Standards for General Hospitals: with a focus on ISO
17090, ISO 22600 and ISO 27799
Yong-En Lee, Korea
12
Non-smoking Webtoon opinion mining, a non-smoking awareness through
the Analysis of posting comments
In Young Choi, Korea
13
Childhood Vaccination Ontology for Social Data Analysis
14
Development of Medical Informatics Principles for Data Repository in
Different Healthcare IT Platforms in Singapore
15
The mediating effect of depression on sleep disorder and fatigue among
reproductive aged women with Psoriasis
Hee Jung Jang, Korea
16
Reliability and Validity of the Persian Version of the Decisional Conflict
Scale in Selecting Mode of Child Delivery
Azam Aslani, Iran
17
Using Simulation to assess user needs for the Point of Care (POC)
Robot development
Mayumi Yoshida, Japan
Jeongah On, Korea
Swee Jin Mok,
Singapore
Hyeong-Suk Lee, Korea
Social Program
Smart Hospital Tour
•
•
Date & Time: Nov. 2 (Wed), 2016, PM 3:30-5:30
Venue: Seoul National University Bundang Hospital [Link]
82, Gumi-ro 173, Bundang-gu, Seongnam, Gyeongi-do, Korea
[ Seoul National University Bundang Hospital ]
Welcome Party
•
•
Date & Time: Nov. 2 (Wed), 2016, PM 6:30-8:30
Venue: Slow Village Brewery and Pub in Yangjae [Link]
67-2, Yangjae-dong, Seocho-gu, Seoul
Gala Dinner
•
•
Date & Time: Nov. 3 (Thu), 2016, PM 6:30-9:30
Venue: Korea House [Link]
10 Toegye-ro 36-gil, Jung-gu, Seoul
[ Korea House ]
Samsung d’light Tour & Seoul City Tour
•
•
Date & Time: Nov. 5 (Sat), 2016, AM 10:00-12:00
Venue: Samsung d’light [Link]
11, Seocho-daero 74-gil, Seocho-gu, Seoul
[ Samsung d'light ]
[ Seoul City Tour: Gangnam ]
18
2016 Conference of
Asia Pacific Association for Medical Informatics
Floor Plan
•
•
•
•
Conference Room: Geomungo A, B, C Hall, Haegeum B Hall
Luncheon: ( 11/3) Geomungo A, B, C Hall
(11/4) Gayageum A, B Hall
VIP & Speaker Room: Cello (3F)
Secretary: Violin (3F)
Geomungo A, B, C Hall (3F) & Haegeum B Hall (3F)
Gayageum A, B Hall (2F)
19
Exhibition
이동형 무대 단
Poster
Session
Geomungo A Hall
Geomungo B Hall
4000 x 1500
4000 x 1500
4000 x 1500
4
3
2
1
Agenda Tower
Registration
4000 x 1500
5
6
7
8
9
10
2000 x
1500
2000 x
1500
2000 x
1500
2000 x
1500
2000 x
1500
2000 x
1500
바이올린 Violin
Secretary
Geomungo C Hall
Water Station
첼로 Cello
VIP & Speaker Room
No.
Exhibitor
No.
Exhibitor
1
HP
6
CHUNGBUK NATIONAL UNIVERSITY
2
BIT computer
7
YONSEI UNIVERSITY
HEALTH SYSTEM
3
ORACLE
8
LifeSemantics
4
LINC (IHE KOREA)
9
ZIKTO
5
SEOUL NATIONAL UNIVERSITY
10
HEALCERION
20
2016 Conference of
Asia Pacific Association for Medical Informatics
Venue Information - Location
The-K hotel Seoul: 70, Baumoe-ro 12-gil, Seocho-gu, Seoul / TEL : 82-2-571-8100
Line 3 Yangjae Station
Sadang-dong
Seocho-gu Office
►
Seoul Arts Center
Seocho-gu Community Center
Shuttle Bus
( to The-K HOTEL SEOUL )
Yangjae
Citizen's Forest
Shuttle Bus ( to The-K HOTEL SEOUL )
The-K Hotel
Patriot Yun Bong-gil
Memorial
SEOUL
Hibrand
Yangjae Flower Market
Gwacheon
Yangjae I.C
►
Bus
Trails
21
Gyeongbu Expressway
Garak-dong ►
Seongnam ►
►
Subway
Yangjae
Citizen's Forest Station
No. 9 Exit at Yangjae Station of Line 3
(Take the shuttle bus in front of Seocho-gu Community Center)
No. 5 Exit at Yangjae Citizen’s Forest Station of Sinbundang Line
(Take the shuttle bus & 5 minutes on foot)
* Line 2 Gangnam, Yangjae Station Sinbundang Line 3 Transit
General blue bus 405A, 405B, 408, 421, 140, 470, 441
No. 10 Exit at Yangjae Station of Line 3
Get off the bus at AT Center, Yangjae Flower Market (10 minutes on foot)
Take the Town bus (NO. 08) at Yangjae Station No. 11 Exit
Yangjae Citizens’ Forest 0.5km (5 minutes)
Yangjaecheon 0.5 km (5 minutes)
Seocho Culture & Art Park 0.2km (3 minutes)
November 3 (Thu), 2016
Oral
Session
1
Session 1: Telemedicine for Global Health Network
• Session 1-1Using eHealth for Addressing Lymphoedema in India
- Arun Rekha Gogia
• Session 1-2Assessing the implementation of the National e-Health Strategic
Planning to support telemedicine projects in Indonesia
- Sumarsono Surahyo, Lutfan Lazuardi
• Session 1-3Informatics beyond boundaries: An experience of technology
enabled remote support in Natural Disasters (Nepal Earthquake
2015 Relief Portal)
- Oommen John, Senthil K. Nachimuthu, Klaus D Veil, S B Gogia,
Vivekanand Jha
• Session 1-4Telemedicine setup and consultation process in a remote health
post in Eastern Nepal - a cross sectional study
- Das RK, Bhattarai A, Gupta PP, Bhatta N
22
2016 Conference of
Asia Pacific Association for Medical Informatics
Using eHealth for Addressing Lymphoedema in India
Arun Rekha Gogia
Chairperson, LE&RN India Chapter
Lymphoedema Consultant and SATHI Member
ABSTRACT
Introduction and Background
ŠŠ
Lymphoedema affects almost 20 million people in
India. Many are ascribed to Filariasis. But despite
adequate control of filariasis due to MDA, lymphoedema, incidence has remained the same even while
other causes like venous disease, cancer and trauma are causing a rise in number of cases. Without
adequate treatment, lymphoedema can have longterm adverse social, economic and psychological
effects, including disability, difficulties with work
and emotional problems.
Lymphoedema is now recognized to be a chronic
disease, no different from Diabetes and Hypertension. So care has to now go beyond hospital based
care. Early diagnosis and care for example at onset,
immediately after surgery or trauma prevents serious problems. But at no stage should the patient be
considered as too late for improvement. Unlike Diabetes, however, knowledge of how to care is missing
not only among patients but also care givers of all
kinds including health professionals of all levels. IT
has a role in Knowledge dissemination, training as
well as monitoring of this chronic problem
There is a dire need to spread awareness that
lymphoedema is both treatable as well as manageable. This has to be done on a mass scale so that
care can be delivered to a large number of patients
quickly.
SATHI – short for Society for Administration of
Telemedicine and Healthcare Informatics has
been propagating use of Telemedicine to promote
healthcare in remote areas. It recently partnered
with Lymphatic Education and research Network,
a US based organization to promote research and
care for Lymphoedema. This work covers aspects
of long term community based group and home
care. This program is aimed at patients, village-level
NGOs and health workers with tele-monitoring and
tele-training after creating the required motivation
physically through advocacy and orientation of the
need for such a care program.
23
Background
ŠŠ
With an estimated 16.7 Million Patients worldwide,
Lymphoedema due to Filariasis is a major health
problem. Of late, thanks to the WHO’s motivated
Program for Elimination of Lymphatic Filariasis
(PELF), the incidence of Filariasis has gone down
in many parts of India. However, the incidence of
Lymphoedema is not decreasing and is not expected to do so at least for the next 10-15 years as that
is related to infection occurring in early childhood. In
addition, many new diseases as well as their treatment, like cancer, trauma, venous disease, recurrent
infections are resulting in fresh cases. The total patient load is estimated at 20 million in India roughly
1-2% of the population.
Methods
ŠŠ
For addressing the public health problem of lymphedema. The major concerns are:
a. Poor availability and accessibility to required
services
b. Poor uptake fo single day mass therapy of DEC/
Albendazole in some areas.
c. Lack of awareness
d. Inadequate capacity for clinical management
e. Inadequate facilities for care
f. Telemedicine is used to improve access to
health services for management of lymphoedema. Alongside the same infrastructure is used
for building capacity and skills of bandaging,
taking measurements etc. Online monitoring is
done as well as discussions on problem cases
with relevant experts.
g. The same centres also provide care for related
common limb problems like, varicose veins, Varicose ulcer and diabetic foot
Results:
ŠŠ
A Centre for Capacity Building for appropriate and
technically sound management of lymphedema is
being created. Through this centre training program
will be carried out for service providers, managers
of the program, application of ICT for management
of lymphoedema etc. It also co-ordinates research
on the most suitable treatment protocols
a. A centre at AIIMS is already working and doing
various trials for materials like bandages garments and pumps in Lymphoedema care.
b. Best use case scenarios can be extrapolated for
community and rural settings where materials
can be provided but training on how to use is required.
c. Remote area use was piloted in a recent camp
in Sitapur District of UP. There were long term
physical gains with tele-monitoring.
d. The same tele system is now being modified for
Online training and further dissemination to patients as well as training of trainers (TOT)
e. A common Database allows sharing and transfer
of patient data.
Conclusions:
ŠŠ
While this is the starting phase of a long term program, we do see hope in long term management of
this chronic problem on a mass scale. IT has a key
role in program implementation
Keywords: Tele-health, Lymphoedema, Tele-monitoring, Volume measurement
24
2016 Conference of
Asia Pacific Association for Medical Informatics
Assessing the implementation of the National e-Health Strategic
Planning to support telemedicine projects in Indonesia
Sumarsono Surahyo,
Lutfan Lazuardi
Faculty of Medicine, Universitas Gadjah Mada,
Yogyakarta, Indonesia
ABSTRACT
e-Health and the urgency to integrate information and
communication technologies (ICT) in national Health
Information Systems and health infrastructure is
growing, and in fact is one of the aim of Indonesia National e-Health Strategic Planning. Based on current
situation in Indonesia, e-Health ecosystem is very
complex and still partial. Most of the e-Health implementations are fragmented and not standardized
which caused difficulties to system interoperability,
consolidation of data (multi-format and standard), and
limited network access to public health information.
The aim of this research is to assess the implementation of the National e-Health Strategic Planning in
supporting healthcare improvement projects in Indonesia. In particular, the research focuses on the telemedicine projects that have been piloted in a number
of regions in Indonesia. Through this project, the
implementation of the National e-Health Strategic
Planning will be evaluated. Factors contributing the
success or failure of telemedicine projects will also
be investigated.
Objectives
ŠŠ
Objectives of this research are to identify how the
National e-Health Strategic Planning affects telemedicine projects as one of the national programs
to improve the health care quality across all regions
in Indonesia, to identify factors contributing the success or failure telemedicine projects in Indonesia
and to identify barriers to widespread adoption of
telemedicine practice in Indonesia.
Methods
ŠŠ
Extensive documents produced in developing National
eHeatlh Strategic Planning and reports from the telemedicine pilot projects were carefully studied and evaluated. Several meetings in National Telemedicine team
in the Ministry of Health Indonesia were conducted, to
produce in-depth understanding the contribution of
Keywords: National eHealth, Telemedicine, assessment
25
how the National eHealth Strategic Planning affected
the telemedicine pilot projects. This study uses the Information Systems Research (ISR) framework to guide
our understanding of the environmental context, which
includes people, organization, and technology.
Results
ŠŠ
Problems discovered in the telemedicine pilot projects were low-technology digital devices (scanner)
which caused high pixel lost and low quality images,
small capacity bandwidth of VPN, and consultation
fee had not been dealt. Meanwhile, there were also
some positive outputs such as centralized server in
the MoH which gave low cost acquisition and maintenance, easy to monitor and easier network re-routing
when there is a change in referral system. Improvement planning has been done by doing assessment
to map infrastructure condition in rural areas, IT human resources, commitment from hospitals and local
governments, survey before installation, and network
fulfilment based on regionalization of referral system.
Conclusion
ŠŠ
The assessment within the field of telemedicine shows
that the program is in its infancy and while its promise is great, the improvement of the program ensure
maximization of benefit. More information on the infrastructure requirements needs to be studied to develop telemedicine solutions. Overall, Indonesia faces
some issues such as underdeveloped infrastructure,
high costs implementation and maintenance, and lack
of technical expertise and motivation from the users.
Meanwhile, legal issues surrounding patient privacy
and confidentiality, and health financial system to enable the sustainability of the program are the next considerations. Lastly, Indonesia has a national agency for
the promotion and development of telemedicine. However, scientific institutions should be more involved
with the development of telemedicine solutions.
Informatics beyond boundaries: An experience of technology
enabled remote support in Natural Disasters (Nepal Earthquake
2015 Relief Portal)
Oommen John1, Senthil K. Nachimuthu2,
Klaus D Veil3, S B Gogia4, Vivekanand Jha5
1_ The George Institute for Global Health India, New Delhi, India
2_ University of Utah School of Medicine, Salt Lake City, Utah, USA
3_ Adj. Associate Professor, Western Sydney University, Sydney, Australia
4_ Asia Pacific Association for Medical Informatics, New Delhi, India
5_ The George Institute for Global Health & International Society of
Nephrology, New Delhi, India
ABSTRACT
Background
ŠŠ
In April 2015, a massive earthquake struck Nepal
leaving over 8,000 people dead and more than
21,000 injured. The major disruption to the healthcare infrastructure in the landlocked mountainous
country added significant stress to the fault lines of
an already fragile healthcare delivery system. Informatics professionals from across the globe quickly
formed a response team which was coordinated by
the leadership and members of the Asia-Pacific Association for Medical Informatics (APAMI) to create
a communication and relief coordination virtual platform to support the health professionals in Nepal at
this time of humanitarian crisis.
Objective
ŠŠ
Testing an online platform for disaster support
Methods
ŠŠ
A web-based electronic communication and project
monitoring tool was deployed using the Sahana platform. The APAMI team set up a dedicated website
and twitter handles with an aim to create awareness
and garner support for the relief work in Nepal. An
online portal in collaboration with the American Nepal Health Association was established. This served
as a tool for the Nepali doctors and healthcare facilities to share information on the materials, medications and other logistic supplies as well as when
and where exactly they were required. Based on this
detailed information volunteers from the Informatics communities then leveraged their affiliations
and partnerships to identify donors and coordinate
the delivery to Nepal through appropriate channels.
One of the specific needs expressed was for clinical
inputs and guidelines on management of medical
emergencies as many specialists were called to
attend to mass causalities. Through the Relief for
Nepal website and twitter accounts, the authors
distributed stepwise evidence based guidelines on
management of crush victims and also as to how to
prevent and manage acute kidney injury based on
the field guide produced by the International Society of Nephrology Renal Disaster Relief Task Force.
This was disseminated through the communication
channels established by APAMI as well as the International Society of Nephrology.
Results
ŠŠ
We describe in this presentation a multi-national
coordinated approach trialed by members of APAMI. We also outline the lessons learned that could
inform key stakeholders on creating virtual communities for disaster relief and humanitarian action
across borders.
Conclusion
ŠŠ
From the 2015 Nepal earthquake relief support
activities we have learned that an informatics community utilizing a real-time information system can
quickly create a support and coordination infrastructure for health professionals working in a disaster-affected country. The communication tools
established were found to be very useful in disseminating knowledge that helped understand the needs
for relief in real-time enabling earlier commencement of life-saving activities. We propose setting up
a task force under APAMI for ‘Health Informatics Beyond Borders’ which could further develop effective
strategies for use of health information technology
in natural disaster relief and supporting health-professionals at the point of need.
Keywords: Natural Disasters, Disaster Relief, Humanitarian relief, Web-based communication and coordination systems, Open Source, APAMI
26
2016 Conference of
Asia Pacific Association for Medical Informatics
Telemedicine setup and consultation process in a remote health
post in Eastern Nepal - a cross sectional study
Das RK1, Bhattarai A2, Gupta PP3, Bhatta N4
ABSTRACT
1_ IT Head, Computer section and Technical Coordinator eHealth and Telemedicine
2_ Technical Coordinator, eHealth and Telemedicine Program
3_ Assistant Professor, Department of General Practice and Emergency Medicine and
Medical Coordinator in eHealth and Telemedicine Program
4_ Prof and Chair, Pulmonology, Critical Care and Sleep Medicine, Department of
Internal Medicine and Focal Person in eHealth and Telemedicine Program,
B. P. Koirala Institute of Health Sciences, Dharan, Nepal
ABSTRACT
Background
ŠŠ
In a country like Nepal where more than half of
populations are in rural areas where the nearest
hospital is situated more than 4 hrs of distance.
The people faces lots of difficulties due to various
factors like provision of timely advice on the health
status, little knowledge or information regarding the
health access. The current hyperbole surrounding
telehealth suggests that it is a revolution in the delivery of health care. The term telehealth as well as
the specific boundaries for possible applications
is difficult to define, as the technological tools and
systems are consistently changing. Telehealth enables continued education, allowing such educated
health professionals, despite location and distance,
to contribute to a better interdisciplinary and overall
expansion of Health care facility. According to Reid,
telehealth means the use of advanced telecommunication technologies to exchange health information
and provide health care services across geographic,
time, social, and cultural barriers.
Methods
ŠŠ
It is a cross-sectional study done in one of the Primary health care centre where we installed and run
Telemedicine program from B.P.Koirala Institute of
Health Sciences. The data was collected from the
records of patient enrolled during the tele consultation process and all statistical tests have significance level of p<0.05. Statistical analysis was performed by using SPSS version 11.0 and confidence
interval analysis for windows.
Keywords: telemedicine, teleconsultation, rural health.
27
Results
ŠŠ
Total 154 patient were enrolled in the study during
8 months of the study begins. There we faced very
difficulties to set up Telemedicine equipments which
include Internet limitations, knowledge of target population and cost of set up. The referral rate was decreased by 42% after starting the tele consultation
process. About 72 % people satisfied with this technique as this saves their time, money, and at the same
time they get consultation by specialist directly.
Conclusion
ŠŠ
Telehealth is a relatively recent and complex technology, which explains why it has not been used to
its full potential, especially in the providing health
care, where it is extremely promising. We are now at
the point where possibilities are clearly seen. Telehealth can be a tool of prime importance to rural
health care deliveries in Nepal. Though, telehealth
is not a magical solution to all the difficulties about
information on health care, but it is a powerful tool
capable of transforming the way with which information is dealt with. It can reduce the number of
referrals and emergency visits of poor people to the
hospitals
November 3 (Thu), 2016
2
Oral
Session
Session 2: Big Data Informatics
•
Session 2-1Data Validation Study for Japanese Sentinel Project
- Rieko Izukura, Takanori Yamashita, Chinatsu Nojiri,
Yasunobu Nohara, Naoki Nakashima
•
Session 2-2A Comprehensive Map of Polygenic Susceptibility Associations
Among Diseases and Traits Based on Large-scale Individual
Genotype Data
-Jung Hun Ohn
•
Session 2-3FCM-EM: Discovering DNA Motifs by Machine Learning
Technique
- Ali Yousefian, Jinwook Choi
•
Session 2-4Development a minimum data set of the information management
system for thalassemia
- Reza Safdari, Marjan Ghazi Saeedi, Seiede Sedigheh Seied
Farajollah, Tayebeh noori, Leila keikha
28
2016 Conference of
Asia Pacific Association for Medical Informatics
Data Validation Study for Japanese Sentinel Project
Rieko Izukura 1, Takanori Yamashita 1, Chinatsu Nojiri 2,
Yasunobu Nohara 1, Naoki Nakashima 1
1_ Medical Information Centre, Kyushu University Hospital, Fukuoka, Japan
2_ Care four, Inc., Fukuoka, Japan
ABSTRACT
Objectives
ŠŠ
Referring to the successful aspect of the Sentinel
Initiative in United States, Medical Information Database Network (MID-NET) project (called as “Japanese
Sentinel Project”) was launched by Japan Ministry
of Health, Labour and Welfare, and Pharmaceuticals
and Medical Devices Agency (PMDA) in 2011. This
project has been aiming to establish new databases
(MID-NET) at 10 selected medical institutions and
to develop a new analysis system, which integrates
electronic medical data from these databases
through the ICT network and analyzes quantitatively
to identify the adverse effects of marketed drugs,
to PMDA, in addition to conventional spontaneous
reporting system. Kyushu University hospital, one of
10 medical institutions, has conducted data validation study with PMDA so far, to prepare the full operation in 2018. In this paper, we examined the scripts
to extract one of the actual drug-induced adverse
events from MID-NET and showed how to assess
data quality.
Methods
ŠŠ
We focused on relative hyperglycemia after the
prescription (Rp.) of Antipsychotic drugs (ADs). The
data collection period was six months and four data
entities (Diagnosis, Prescription, laboratory test, and
Medical reimbursement information) were used to
describe the scripts on MID-NET. Subjects, who were
already diabetic before ADs Rp, were previously excluded by using the diagnosis of diabetes mellitus (DM),
the prescription of anti-diabetic drugs, and DM-related nutritional education program (Script 0: n=2129).
Then Script 0 was each combined with 3 keys of hyperglycemia: DM diagnosis (Diagnosis: Script1), anti-diabetic drugs Rp. (Prescription: Script2) or high blood
glycose (HBG) (Results of laboratory test: Script3).
HBG, a criterion to detect the relative hyperglycemia,
was defined as “<140 mg/dL before and >199 mg/dL
after ADs Rp.” by DM specialist. The data from single or
combined scripts were counted and calculated coincidence rates.
Results
ŠŠ
The smallest number of patients was by Results
of laboratory test at Script 3 (n=32, 1.5%). However we found 1086 patients (51%) out of 2129
patients did not receive BG test after ADs Rp. and
some of them probably have hidden hyperglycemia.
Both combination of Script 1 & Script 3 and Script
2 & Script 3 were low coincidence rate (0.2%, 0.5%,
respectively), which would be reflected that HBG
were not necessarily after diagnosed DM or/and prescribed DM drugs.
Conclusions
ŠŠ
Since a script used Results of laboratory test alone
is difficult to identify hyperglycemia as an adverse
effect of ADs Rp. The combination of scripts, which
connects scripts with “OR”, but not with “AND”, is
able to identify more hyperglycemic subjects (5.2%),
would be desirable to improve a level of sensitivity.
Additionally, it was indicated that MID-NET will make
it possible to detect such an adverse event with less
laboratory test or with less data collection.
Table1 The coincidence ratio of MID-NET data
n (%)
Script 0
Script 0
Script3 HBG+
Script 1 or Script 2 or Script 3
BG test procedure -
2129
32 (1.5)
110 (5.2)
1043 (49)
Script 1 DM diagnosis+
50 (2.3)
5 (0.2)
Script 2 DM drug Rp.+
66 (3.1)
10 (0.5)
Script 3 HBG+
32 (1.5)
-
Keywords: MID-NET, adverse effects, big data, drug exposure
29
A Comprehensive Map of Polygenic Susceptibility Associations
Among Diseases and Traits Based on Large-scale Individual
Genotype Data
Jung Hun Ohn
ABSTRACT
Objective
ŠŠ
The aim is to explore the landscape of diseases or
traits with respect to genetic risk correlations to
find clinically relevant associations.
Methods
ŠŠ
Based on 2,504 individual genotype data and database of genetic variants linked with human diseases
or traits, genetic risk scores (GRS) of diseases or
traits are calculated for each individual. Association
between traits and diseases is statistically evaluated by pairwise correlation analysis of GRS. The significant associations are compared to disease comorbidity correlations obtained from hospital claims
from more than 30 million patients in US.
Results
ŠŠ
Correlation analysis of GRS reveals 823 significant
correlations among 78 diseases or 89 traits (q
value < 0.01). It reveals the existence of common
genetic backgrounds between diseases or traits,
such as atrial fibrillation vs. ischemic stroke (r=0.59),
venous thromboembolism vs. susceptibility for severe malaria (r=0.40), coronary heart disease vs.
open angle glaucoma (r=0.29), type 1 diabetes vs.
vitiligo (r=0.31), type 2 diabetes vs. low birth weight
(r=0.26), and Helicobacter pylori infection vs. self-reported allergy (r=-0.28). Significant polygenic risk
correlations are found in 464 (56.4%) associations,
especially among autoimmune diseases, despite the
absence of shared genetic variants or in linkage disequilibrium. When 312 significant correlations between diseases are mapped to disease comorbidity
correlations from Medicare claims data from the US,
108 (34.6%) polygenic susceptibility correlations
are validated to have epidemiologic evidence, such
as coronary heart disease vs. intracranial aneurysm
(r = 0.26 and relative risk, 1.57 [99% CI 1.56-1.58]),
asthma vs. obesity (r = 0.07 and relative risk, 2.18
[99% CI 2.14-2.22]), and open angle glaucoma vs.
coronary heart disease (r = 0.29 and relative risk,
1.08 [99% CI 1.07-1.09]).
Conclusions
ŠŠ
The study suggests that common genetic background exists between comorbid diseases or traits.
The polygenic risk correlation approach provides
candidate disease comorbidity associations warranting further epidemiologic investigations.
Keywords: genetic risk score; disease network; the 1000 Genomes Project
30
2016 Conference of
Asia Pacific Association for Medical Informatics
FCM-EM: Discovering DNA Motifs by Machine
Learning Techniques
Ali Yousefian 1, Jinwook Choi 2
1_ Interdisciplinary Program in Bioengineering, Graduate
School, Seoul National University
2_ Dept. of Biomedical Engineering, College of Medicine,
Seoul National University
ABSTRACT
This study aims to develop a framework for discovering DNA motifs, where fuzzy C-means (FCM)
membership functions, and an EM technique are
employed to extract putative motifs allowing Indels.
We present FCM-EM, a motif discovery algorithm
designed to find DNA-binding motifs in ChIP-Seq
data. The discovered motifs in ChIP-Seq dataset are
matched to known motifs using TOMTOM and its
calculated E-value shown the statistical significance
of each discovered motif as well.
Objective
ŠŠ
Understanding genetic sequences is one of the fundamental tasks of health advancements due to the
high correlation of genes with diseases and drugs.
An important problem within genetic sequence understanding is related to transcription factors (TFs),
which are regulatory proteins that bind to DNA.
Each different TF binds to specific transcription
factor binding sites (TFBSs) on the DNA sequence
to regulate cell machinery. In this work, we focus on
accurately classifying and understanding the DNA
subsequences that TFs bind to, which will allow us
to better understand the underlying biological processes and potentially influence biomedical studies
of human health.
Methodology
ŠŠ
However, MEME scales poorly with large datasets
and shows inability to find the motifs including in-
Keywords: Motif discovery, fuzzy C-mean, expectation maximization.
31
sertions and deletions. This study aims to develop
a framework for discovering DNA motifs, where
fuzzy C-means (FCM) membership functions, and
an EM technique are employed to extract putative
motifs allowing Indels. The method proposed in the
study follows the following steps: (a) clustering subsequences into a certain number of clusters using
FCM (b) utilization of fuzzy membership values of
each subsequence as an initial posterior probability
values in EM technique (c) testing each PWM, clusters center, to see whether it is statistically meaningful or not.
Experimental Results
ŠŠ
To evaluate the proposed algorithm, we use ChIPSeq data generated by the Myers Lab at the Hudson
Alpha Institute for Biotechnology. The discovered
motifs are aligned and matched to known motifs
using a tool called TOMTOM and the significance of
results are considered by calculated E-value of each
discovered motif as well.
Conclusion
ŠŠ
In this study, employing FCM clustering algorithm
for initializing EM technique, has been proposed to
provide a novel tool to find the motifs in DNA sequences. Unlike MEME, this algorithm is able to find
the motifs with Indels, and EM algorithm can also
converge in the first few iterations as well.
Development a minimum data set of the information management
system for thalassemia
Reza Safdari 1, Marjan Ghazi Saeedi 2, Seiede Sedigheh
Seied Farajollah 3, Tayebeh noori 3, Leila keikha 4*
1_ Professor, Department of Health Information Management, School Allied Medical
Sciences, Tehran University of Medical Sciences, Tehran, Iran.
2_ Assistant Professor, Department of Health Information Management, School
Allied Medical Sciences, Tehran University of Medical Sciences, Tehran, Iran.
3_ PhD student of Health Information Management, Department of Health Information
Management, School Allied Medical Sciences, Tehran University of Medical
Sciences, Tehran, Iran.
4_ PhD student of Health Information Management, Department of Health Information
Management, School Allied Medical Sciences, Tehran University of Medical
Sciences, Tehran, Iran. Department of Health Information Technologies, Paramedical
School, Zahedan University of Medical Sciences, Zahedan, Iran
*_ p
resenter and correspond Author
ABSTRACT
Thalassemia as the most common chronic genetic
disorder in 60 countries, affecting lives of about
10.000 people annually. Also, in Iran the thalassemia disease recognized as the most important genetic disorder. But there is not a minimum data set
to management thalassemia disease. The aim of the
study was to develop and create a minimum data set
on the information system of thalassemia for Iran.
Descriptive, cross-sectional study was conducted in
2015.
Material and methods
ŠŠ
Data was collected from patient medical records
of ALI ASGHAR special hospital affiliated with
Zahedan University of medical sciences. Moreover resources were included in research such as,
guidelines about thalassemia, articles, books and
forms which supplemented on hematology books.
Different medical records in D-56 category related
to thalassemia based on international classification
of diseases 10th revision (ICD-10) were randomly
selected. In order to extract the data element from
the source listed above, a form was used. Collected
data were divided into clinical and administrative
categories using a checklist. Then, in order to make
consensus about collected data, the Decision Delphi
technique was used via a questionnaire. The content
validity and reliability of questionnaire were evaluated by expert’s viewpoint and test-retest method,
respectively.
Results
ŠŠ
MDS of thalassemia was developed. This MDS divided into two categories: administrative and clinical
with 5 and 17 sections were 157 and 284 data elements, respectively.
Conclusion
ŠŠ
this study showed that uniform data element about
thalassemia do not create in Iran. Thus, a MDS was
designed for thalassemia in Iran. Development of
thalassemia MDS reinforces the interoperability
of data and information of thalassemia disease
throughout registry system, health information
management system and electronic health records
(HER). Moreover, it can be used by reimbursement
for pay treatment cost and claim. Information provided from MDS creates valuable resources for
assessment, treatment planning, continuous evaluation of patient progress and performance. In addition to, establishment of MDS enables comparability
data which extracted from different sources which
results in the improvement of the quality care.
Keywords: thalassemia, minimum data set, information management system
32
November 3 (Thu), 2016
3
Oral
Session
Session 3: Health Data Exchange
33
•
Session 3-1Implementation of standards-based, vendor neutral archivedriven API platform and client
- Jung Bin Kim, Dong Ouk Kim, Yong Seok Kim
•
Session 3-2Unlocking Interoperability Puzzle with the Global Language of
Health - SNOMED CT
- Dr Michael Bainbridge
•
Session 3-3Pilot study for a mapping from HL7/SS-MIX2 to CDISC SDTM
- Naoki Nishimoto, Jun Kunikata, Koji Uemura, Sumiko Akahori,
Tomoaki Sogo, Yuma Tani, Hideto Yokoi
•
Session 3-4Data Quality Management for Clinical Contents Model
- Yookyung Boo, SoJeong Han, JaeWoo Lee, YoungTaeq Koh, and
KyungHee Cho
Implementation of standards-based, vendor neutral
archive-driven API platform and client
Jung Bin Kim, Dong Ouk
Kim, Yong Seok Kim
R&D Center, INFINITT Healthcare Co.,Ltd. Seoul
ABSTRACT
Objective
ŠŠ
This study is aimed at implementing a vendor neutral
archive (VNA)-based platform for medical images
that enables the use and consumption of API. The
VNA to be developed uses various profiles suggested by Integrating the Healthcare Enterprise (IHE)
and connects them to operate as a standards-based
API platform. Such API platform-related technologies can further be applied to VNA in creating an
ecosystem for medical images. The objectives of
this study include: i) to implement resources of IHE
XDS.b, IHE MHD, and HL7 FHIR to a VNA and apply
OpenAuth to the VNA for constructing a platform
for medical image data. ii) to allow users to obtain
right to access the platform, consume resources,
and publish medical images and documents through
an OpenAuth-based client application. iii) to verify
that the entire system is operable through the built
platform and the client application that uses the
platform.
Methods
ŠŠ
Metadata can be input to the medical images and
documents via IHE XDS.b. This enables the archiving
and viewing of the medical data in a patient-centric
way. IHE MHD, which allows XDS.b to interface with
Representational State Transfer (REST), enables
querying, retrieving, and displaying of data in even
low-specification devices such as mobile gadgets.
The platform was constructed to enable users to
permit designated client applications to access the
data in the VNA and to upload medical data, by applying OpenAuth 2.0.
Results
ŠŠ
To realize the system, an IHE MHD-based scenario
and a SMART on FHIR-based scenario were combined. IHE MHD is defined to be able to provide a
FHIR bundle and consume FHIR resources, when
combined with IHE IUA. FHIR provides atomic resource objects rather than composite ones. SMART
on FHIR suggests that scope management should
be done to allow access to each resource via OpenAuth. To verify if the system is operable when the
two scenarios are combined, Healthcare Content
Uploader, a client application that supports OpenAuth, FHIR, and MHD was developed.
Conclusions
ŠŠ
The purpose of this study is to develop a VNA-driven
API platform that is easy-to-use and secure, by applying MHD that can provide SOAP-based XDS.b via
RESTful; FHIR resources as suggested by SMART
on FHIR; and OpenAuth. The study shows that it is
possible for third party applications to use API of the
VNA based on the system. The results verifies that
VNA resources can be shared across different applications via security measures; non-DICOM objects
can be saved and managed; and HIE can be further
achieved. It is our expectation that this technology
can enhance existing PACS to realize an API platform through medical image-based VNAs and build
health IT ecosystems based on API platforms in
which users can develop and use various healthcare
applications.
Keywords: Vendor Neutral Archive, Application Programming Interface, Healthcare IT Platform, Healthcare IT Standards, Healthcare IT Security
34
2016 Conference of
Asia Pacific Association for Medical Informatics
Unlocking Interoperability Puzzle with the
Global Language of Health - SNOMED CT
Dr Michael Bainbridge
Clinical Engagement Lead (APAC) on behalf of
International Health Terminology Standards
Development Organisation (IHTSDO)
ABSTRACT
Introduction
ŠŠ
In healthcare, interoperability and exchange of data
is facilitated by ensuring that different information
systems and software applications communicate,
exchange data, and are able to use the information
that has been exchanged. However, the healthcare
sectors across the world are faced with a mixed
economy of paper based medical records, local Document Management Systems, Electronic Medical
Records and other data capture and exchange formats. Healthcare providers are attempting to deliver
care through summarising patients’ medical records
using these various formats and different approaches, including free text (using regional dialect and
acronyms), local and proprietary terminologies, classifications and other coding systems designed for
reporting and reimbursement purposes. Proliferation
of terminologies and coding schemes increase the
number of errors and hinder clinical data exchange
and meaningful use. In order to improve patient care
and safety both at individual and national levels, patient data needs to be captured accurately, shared
effectively, and its meaning interpreted consistently. While the use of Electronic Health Records has
enabled healthcare providers to capture and share
clinical data effectively within a single system, it is
important that different clinical systems are also
able to communicate and understand each other.
The only way to safely achieve this shared understanding is for clinical systems to speak a common
language. A clinical terminology, such as SNOMED
CT, provides this common language for healthcare
systems and providers to communicate in this way.
In addition, however, it also enables shared clinical
data to be effectively utilized for clinical research,
analytics and decision support and multiple other
secondary uses, which contribute to improvements
in healthcare delivery.
35
Healthcare organisations and governments across
the world are turning their attention to semantic
interoperability – the ability of two or more systems
to exchange information and to use the information that has been exchanged. As large volumes of
information (often referred to as big data) become
available as a result of the increased use of Health
Information Exchange systems, so much of this
information will have little value if it cannot be processed and analysed dependably. SNOMED CT
provides a standardised representation for clinical
phrases and concepts captured by clinicians, in a
way that supports the electronic interpretation of
their meaning. SNOMED CT enables clinical information to be represented consistently, reliably and
at a level of granularity required by clinicians, while
also enabling clinical data aggregation, clinical research, data analytics and decision support.
SNOMED CT is the most comprehensive, multilingual clinical healthcare terminology in the world. It
is used in over 50 countries including the US, UK,
Sweden, Spain, India, Hong Kong (P.R.C), Malaysia,
Australia, Canada and many others. SNOMED CT
supports the representation of high quality clinical
content in digital health records. The implementation
of SNOMED CT enables a broad range of benefits to
a wide variety of stakeholders – including healthcare
providers, consumers and administrators, medical
researchers and entire populations.
In this presentation, we will discuss the key features
of SNOMED CT and the significant benefits that
can be achieved through its implementation. We also
consider both the key determinants for effective
benefit realization, and how the increasing utilization
of SNOMED CT’s power can increase the benefits
achieved.
Presentation Brief
ŠŠ
SNOMED CT provides a number of features, which
can be used in the implementation of health records,
including:
a. uniquely identified meanings (i.e. concepts);
b. human-readable labels (i.e. descriptions);
c. specialization hierarchies and formal description-logic based definitions (i.e. relationships); and
d. the ability to customize the terminology, by
adding content, language preferences, subsets,
translations, mappings and other implementation-specific requirements (i.e. extensions and
reference sets).
Conclusion
ŠŠ
SNOMED CT contributes to the improvement of
patient care by supporting the development of high
quality, clinical content in digital health records.
SNOMED CT supports semantic interoperability,
which improves quality, safety, efficiency, and efficacy of healthcare delivery. The range and depth
of benefits achieved through the implementation
of SNOMED CT is determined by a number of factors, including the design decisions that are made.
Through considered design and an increasing utilization of SNOMED CT’s power, the effective delivery
of these benefits can be further enhanced.
However, not all of these features need to be implemented in order to achieve benefits from SNOMED
CT. Systems that only require a few of these features to be implemented using a simple approach,
still benefit from the opportunity to exchange data
using a common language. Similarly, while SNOMED
CT spans a wide range of clinical topics and specialties, systems can achieve a variety of substantial
benefits using only a small subset of this content.
A number of different strategies may be used to deliver benefits with SNOMED CT. These range from
an incremental approach, which starts with a simple
subset of SNOMED CT’s features and evolves to
meet emergent requirements, through to planning
upfront to achieve the depth of benefits enabled by
a sophisticated implementation.
This presentation draws upon the work done by the
IHTSDO on SNOMED CT benefits realization and
the general business case for adoption of SNOMED
CT to achieve semantic interoperability. Some of
the key areas covered by the presentation include
SNOMED CT use in data analytics and decision support, and how it contributes to achieving the benefit
realization.
36
2016 Conference of
Asia Pacific Association for Medical Informatics
Pilot study for a mapping from HL7/SS-MIX2 to CDISC SDTM
Naoki Nishimoto1, Jun Kunikata1, Koji Uemura2,
Sumiko Akahori2, Tomoaki Sogo2, Yuma Tani3, Hideto Yokoi1,2
1_ Clinical Research Support Center, Kagawa University Hospital, Kita-gun, Japan
2_ Department of Medical Informatics, Kagawa University Hospital, Kita-gun, Japan
3_ Department of Medical Informatics, Graduate School of Medicine, Kita-gun, Japan
ABSTRACT
Objectives
ŠŠ
The purpose of the study was to explore the exchangeability between HL7/SS-MIX2 and CDISC
SDTM (Clinical Data Interchange Standard Consortium, Study Data Tabulation Model) formatted
data for laboratory test results. And the secondary
purpose was to explore evaluation framework and to
evaluate mapping precision.
Methods
ŠŠ
We mapped HL7 field name to CDISC SDTM variable name. LB (Laboratory test results) domain
was selected because the HL7 variable name of
laboratory test has less variability than others,
such as disease name or past histories. To test the
exchangeability, we accumulated 77 cases over
330 subarachnoid hemorrhage patients from EMR
in Kagawa university hospital. The test data were
obtained from April 1, 2014-March 31, 2015.
Bayesian sample size estimation was conducted to
construct the evaluation framework (J Whitehead,
2008). We set Un as a mapping proportion,
η=0.95, ζ=0.90.
P ( p ≥ P o| y ) ≥ η
P ( p < p *| U n – 1 ) ≥ ζ
We applied a Bayesian method to estimate the
mapped count using simple Binomial prior distribution. MCMC procedure, which uses Markov-Chain
Monte Carlo technique, was used in SAS 9.4 (SAS
Institute Inc.).
Results
ŠŠ
LB domain contains 45 variables as it is, we identified 11 one-to-one variables (11/45=24%), such as
LBORRES(results in original unit). And LBORRESU,
which contains unit of test result, required variable
transformation dividing small portion of variable
name because the data was stored in HL7 format
like “g/dl^g/dl^99Z15.” 12 out of 45 variables (27%)
were seen as this category. Other variables could
not be transformed because they were specific
to a clinical trial. These type of variable should be
transformed with encoding the standard code set
of JLAC10 or LOINC code. Interim structure for
transformation was inspired. 108 mapped count
over 193 cases were estimated for the sample size
calculation in evaluation framework. In this study,
we used only variable names for mapping, mapped
proportion was estimated using Bayesian technique.
We obtained 10.83 one-to-one variable and its 95%
highest posterior density credible (HPD) interval was
5-16 variables.
Conclusion
ŠŠ
Limiting the LB domain in CDISC SDTM, we obtained only 24% mapped variable from HL7 field
name. Mapping precision had a wider 95% HPD interval. The result of this pilot study indicated that refined mapping algorithm was required for the future
development research. We are trying the contents
specific mapping from 77 cases.
Keywords: CDISC SDTM, LB domain, HL7, mapping, Bayesian estimation
37
Data Quality Management for Clinical Contents Model
Yookyung Boo1, SoJeong Han2, JaeWoo Lee3,4,
YoungTaeq Koh5, and KyungHee Cho3
1_ Department of Healthcare Management, Eulji University,
Seongnam-si, Republic of Korea
2_ Department of Healthcare Management, Graduate School,
Eulji University, Seongnam-si, Republic of Korea
3_ Department of Family Medicine, National Health Insurance
Service Ilsan Hospital, Goyang-si, Republic of Korea
4_ Department of Policy Research Affair, National Health Insurance
Service Ilsan Hospital, Goyang-si, Republic of Korea
5_ Department of General Surgery, National Police Hostipal,
Seoul-si, Republic of Korea
ABSTRACT
Objectives
ŠŠ
Healthcare system’s interoperability is one of the
most important goals for meaningful use of the
Electronic Health Records (EHR). It is essential to
facilitate IT support for workflow management, decision support systems, and evidence-based healthcare, as well as the secondary use of EHR across
healthcare organizations. In Korea, the researches
on structured data elements in EHRs have been
progressed since 2008. The clinical contents model
(CCM) is an information model designed to represent clinical information in EHR systems across
organizations. In order to ensure EHR system’s semantic interoperability in clinical information, the
development principle for clinical information model
is required to reflect its objective and function. Data
validation is often a topic of great importance when
it comes to databases. By practicing simple data
validation rules, databases become more consistent,
and functional providing more values to their users.
This research performed data quality control on previously developed CCM for medical facilities’ users
before CCM loading is done on CCM platform. In the
research, the data quality control strategy is investigated for CCM, which is continuously usable in the
future, based on the results of CCM’s data quality
control.
Methods
ŠŠ
The subject of the research is the clinical contents
models (CCM) developed between 2008, and 2015.
CCM data quality control tools such as guideline for
CCM development, and CCM instance review form
developed by EHR organization in 2009 are used.
The reviews are composed of 22 checkpoints in 8
categories. The reviews are randomly sampled in 5
domains of CCM, and evaluated by 5 health information professionals who worked more than 3 years in
EMR document development to ensure its suitability
as assessment tool.
Results
ŠŠ
After evaluating 452 CCMs in total, the significant
improvements seem to be required. For naming convention, common expressions are generally followed
except for some case sensitivity, typing errors, and
etc. In case of CCM model, it is found that the laboratory domain is developed in E-V platform instead
of E-Q-V platform. Common terminology system for
mapping of entity, qualifier, value set, and value usually assign codes such as SNOMED CT, UMLS, Korean Standard Terminology (KOSTOM), Korean Drug
Code (KDC). However, local codes existed in some
cases. In case of evaluation scale tool, metadata elements like development main body, revision year, and
etc were omitted. In addition, considerable number
of omission of the information for CCM development
main body, and description on cases of application is
found. CCMs developed after 2012 omitted all the
expressions of markup language such as CCML.
Conclusion
ŠŠ
Considerable change and supplement on review object are required, and each CCM should be reviewed
before publication. Inorder to support semantic
interoperability, clinical contents model makes the
standard terminology system mapping as a principle
regarding all the elements such as entity, qualifier,
value, and etc. In general, KOSTOM is designed to
map with other terminology systems such as UMLS,
38
2016 Conference of
Asia Pacific Association for Medical Informatics
and SNOMED CT when some concepts are missing.
However, some CCMs require mapping standard
terminology system as they are only assigned of
local code. In addition, there are some cases where
CCM development principle, and methods are not
followed as CCM is developed as annual research
project by different organizations since 2012 after the establishment by Center for EHR between
2008 and 2010. This suggests that the verification
and evaluation of developed CCM; and observance
of CCM development principle were missing due to
absence of governing body. Therefore, management
system is required in order to promote use in medical facilities, and continuous development and quality control for CCM.
Table1 Current state of CCM development, and number of sampling
CCM development year
Number of CCM
development
Target Domain
Number of Random
sampling
2012
Clinical observation , Procedure,
Medication & Laboratory
Clinical observation
114
11
2013
Clinical observation & Laboratory
334
33
2014
Laboratory
106
10
2015
Clinical observation & Laboratory
713
71
4,543
452
2008 - 2010
Total
3,276
327
Table2 Composition of CCM instance review form
Category
Description
To check whether the CCM
Naming Convention Filename complies with the
Naming Convention
Check points
1. CCM filename and Entity Name match up
2. Does CCM filename match up with Value representation?
3. Is Panel, and Cluster used appropriately in the CCM Filename?
CCM Model
To check whether the CCM
complies with the E-Q-V
model
1. Does CCM instance follow the formal structure of the CCM model,
which is Entity – Attribute (Qualifier/Modifier) – Value or Entity – Value
model?
Entity
To check whether the entity
complies with the rules of
CCM model
1. Does Entity Name comply with the value representation?
2. Does Entity Name belong to appropriate Entity Domain?
3. Is Entity Name mapped to a standard terminology system?
Qualifier
To check whether the qualifier complies with the rules of
CCM model
1. Does Qualifier Name comply with the Value representation?
2. Does Qualifier Name belong to qualifier set according to the Entity
domain?
3. Does Qualifier comply with atomicity?
4. Is Qualifier Name mapped to a standard terminology system?
Modifier
To check the appropriateness
of using Modifier
1. Subject
2. Negation
3. Uncertainty
Value
To check whether the Value
complies with the rules of
CCM model
1. Does the Value Name comply with the Value representation?
2. Do Qualifier Name, and Value Set match up?
3. Is Value entirely composed of those of the same nature?
4. Is Value Name mapped to a standard terminology system?
Panel
To check whether the Panel
complies with the rules of
CCM model
1. I s CCM Panel instance composed of items having the same datatype
eValue of two or more?
2. Does CCM Panel Instance have the same qualifiers?
Cluster
To check whether the
Cluster complies with the
rules of CCM model
1. Is CCM Cluster instance composed of items having a dataype with
2 or more of the other eValue, and does the CCM Cluster instance
inclue more than 1 Panel?
Data Type Use
Case
To check the compliance of
standard data type
1. Complying data type of HL7 V3
Metadata
Inclusion of metadata
1. I n order to continuously manage the changes in model,
it has metadata information.
CCML
Definition of marking language
1. All the CCM is marked as CCML.
Keywords: structured data element, clinical contents model, electronic health records, semantic interoperability, data quality control
39
4
November 3 (Thu), 2016
Oral
Session
Session 4: Nursing Informatics
•
Session 4-1Develop and validate an empowerment tool for pregnant women
- A. Aslani, F. Erfanian, S. Yaghoutikhorasani, A. Dadashi, R. Rezaee,
•
Session 4-2The development and acceptability evaluation of the virtual
simulated hospital training program for nursing students
- Yanling Wang, Fangqin Wu, Meihua Ji, Ying Wu,
•
Session 4-3The need Assessment on Long-term Care Information In Taiwan
- Abdallah Ahmed Elbakkoush, Suleman Atique, Chien-Tsai Liu
40
2016 Conference of
Asia Pacific Association for Medical Informatics
Develop and validate an empowerment tool for pregnant women
A. Aslani1, F. Erfanian2, S. Yaghoutikhorasani2,
A. Dadashi3, R. Rezaee1,
1_ Health Human Resources Research Center, School of
Management & Information Sciences, Shiraz University of
Medical Sciences, Shiraz, Iran.
2_ Nursing and Midwifery School, Mashhad University of Medical
Sciences, Mashhad, Iran.
3_ Department of Medical Informatics, Mashhad University of
Medical Sciences, Mashhad, Iran.
ABSTRACT
Objective
ŠŠ
Findings
ŠŠ
Method
ŠŠ
Conclusion
ŠŠ
Women empowerment especially pregnant women
are a basic approach to improve women’s health status and reduce complications during pregnancy. The
“right to health” cannot be achieved only through
direct service, though the aim of this study was to
develop and validate an empowerment tool for pregnant women which is applicable in prenatal training.
To develop the tool five steps were taken: literature
review, generating item pool, content validity testing,
administration of draft scale and finally explanatory
factor analysis. The appropriate and final version of
the tool with 14 items administered to 125 pregnant women undergoing prenatal visits at 4 public
health centers in Mashhad.
Keywords: prenatal training, pregnancy, empowerment, evaluation
41
Item-total and subscale-total scores correlated
significantly with Cronbach’s alpha coefficients of
0.71 and 0.73-0.76 respectively, which confirmed
the internal consistency of the tool. Evaluation of
construct validity through factor analysis revealed
8 items adapted to three factors: self-esteem, optimism over the future, and self-efficacy.
The established tool is a valid and reliable tool with
10 items for evaluating prenatal training in pregnant
women.
The development and acceptability evaluation of the virtual simulated hospital training program for nursing students
Yanling Wang, Fangqin Wu,
Meihua Ji, Ying Wu,
School of Nursing, Capital Medical University,
Beijing, China (# co-first author)
ABSTRACT
Objectives
ŠŠ
Nursing educators often face a dilemma of assigning
inexperienced nursing students to a real clinical setting in which skilled techniques and solid knowledge
base are required in providing safe patient care. The
training in a skills lab is not adequate in addressing
this need and tends to be attentive to students in
developing practical skills and abilities, and does not
fully reflect real clinical scenarios in training and promoting clinical thinking as well as other higher level
nursing competencies. Together with limited clinical
resources and high demands in patient care, there is
a need in developing a training program which could
adequately prepare nursing students in improving
their clinical competencies without violation of patient’s safety before they enter a real clinical setting.
The purpose of this study is to develop a virtual
simulated hospital training program and evaluate its
technological acceptability and usability among student nurses.
Methods
ŠŠ
This study includes two phases: development of the
virtual hospital and evaluation of its acceptability
and usability. The virtual hospital is structured via
a game engine software named “Unity 3D”. A modified acceptance evaluation questionnaire based on
Technology Acceptance Model (TAM) was used to
test the usability of the program, which includes four
domains: Perceived Usefulness (PU), Perceived Ease
of Use (PEOU), Attitudes Towards Usage (ATU) and
Behavioral Intention to Use (BIU), with 41 items and
rated with 5-point Likert scale.
intervention, daily nursing care, and communication.
Interactive interfaces of the training program are
shown in Figure 1. A total of 129 student nurses
tried the program and completed the questionnaire. The means of the four domains regarding to
the system are: PU: 4.02-4.16, PEOU: 3.83~4.19,
ATU: 3.81~4.22, BIU: 4.08~4.26. The Cronbach’s α
of each domain is 0.967, 0.967, 0.929, 0.972 respectively. The average scores of items perceived to
improve clinical nursing competencies were scored
higher than those perceived to help students building
nursing knowledge and practical skills. The students
considered that the virtual hospital training program
could provide adequate training in following physician
orders, improving clinical thinking, and conducting
health observations, communications and so on.
Conclusion
ŠŠ
This training program has easy to use interfaces, the
mean scores of the four domains indicated that this
is a useful training tool in developing clinical nursing
competences for nursing students before they enter
a real clinical setting, and it is user friendly and pertained to improve clinical thinking competencies as
major objectives in using this training program.
Figure 1 Interactive interfaces
Results
ŠŠ
Nursing tasks reflecting different levels of nursing
competencies within this interactive virtual hospital include health assessment, observation, clinical
analysis and decision making, prioritization, nursing
Keywords: Virtual simulated training; technology acceptance model (TAM); patient’s safety
42
2016 Conference of
Asia Pacific Association for Medical Informatics
The need Assessment on Long-term Care Information In Taiwan
Abdallah Ahmed Elbakkoush,
Suleman Atique, Chien-Tsai Liu
Graduate Institute of Biomedical Informatics,
Taipei Medical University, Taiwan
ABSTRACT
Objective
ŠŠ
This study aimed to investigate the information
needs along with future challenges and recommendations for long-term care in long-term care institution residents and their families in Taiwan. We want
to find out the method of information retrieval adopted by population regarding Long Term Care
Methods
ŠŠ
It is a survey based study which involved convenience sampling. We developed a questionnaire to
investigate the state of long term care information
services in Taiwan. The questionnaire mainly included questions regarding Long Term Care needs Requirements and Long Term care funding experience
and satisfaction information regarding information
sources. For this we used 5 point Likert Scale from
‘Very Satisfied’ to ‘Very Dissatisfied’. Another important component consisted of demographic information like Age, Gender, educational level and occupation. It describes public long term care services and
information needs. To enhance the credibility of our
findings we used Chi-square test to find out difference between demographic characteristics and long
term care needs.
search was the most popular method to gain longterm care information, although the government had
built websites to provide information, both the utilization rate and satisfaction regarding these portals
was low. Public mainly uses websites as source of
long-term care issues information network (54.6%)
and television news (48.8%), followed by the information leaflets (27.3%) and social media (18.5%),
through friends and relatives, and apparent network
TV multimedia is mainly transmitted by long-term
care information interface, most people are mostly from the Internet to reach the issues related to
long-term care; fewer people use leaflets, radio (both
6.2%) and telephone (5.4%), other sources include
information retrieval by individuals through insurance
agents, and related organizations or associations
who actually work in long-term care services. Table
2 shows the information source of long term care
among Taiwanese population.
Conclusion
ŠŠ
To improve the policy acceptance and fulfill the information need of general population, it would be necessary to refine and improve the websites providing
information regarding long term care in Taiwan.
Results
ŠŠ
A total of 265 questionnaires were collected including 201 copies of manual questionnaires and 64
online questionnaires. Among them 5 were invalid
questionnaires and 260 valid questionnaires. Table
1 shows the demographic characteristics of the
survey participants. Results showed that website
Keywords: Long-term Care, information need assessment, questionnaire survey
43
Table1 Basic Demographic Characteristics
Feature
Table2 Main Source of Information Retrieval
Frequency
Percentage
Information Source
Frequency
Percentage
20-29
73
28.1%
Web Search
200
76.9%
30-39
62
23.9%
Word of mouth
86
33.1%
40-49
47
18.1%
Newspapers and magazines
79
30.4%
50-59
51
19.6%
Over 60 years
Community Media
60
23.1%
27
10.4%
Advisory long-term care services
55
21.2%
Primary
7
2.7%
City and county long-term care
management consulting center
50
19.2%
Secondary
12
4.6%
Information leaflets
22
8.5%
High school level
40
15.4%
Graduate
144
55.4%
Advisory long-term care associations
22
8.5%
Post Graduate
57
21.9%
others
1
0.4%
Class engineering and manufac59
turing
22.7%
Free services
53
20.4%
Medical long-term care industry 47
18.1%
At home
39
15.0%
student
30
11.5%
Military and government
24
9.2%
Finance and insurance
8
3.1%
Age
Educational Level
Occupation
44
November 3 (Thu), 2016
5
Oral
Session
Session 5: P
ersonal Health Record (PHR) and
Hospital SystemHospital System
45
•
Session 5-1Development of Personal Health Records Architecture
for Community Oriented Holistic Care to be in Line with
International Standards
- Heng-Shuen Chen, Ching-Chu Liao, Pei-Shan Yang,
Yu-Yuan Yu, Hung-Pin Hou, Heng-Chang Chen
•
Session 5-2Development of a Mobile Application in Conjunction with a Web
Service for Personal Health Records regarding Atopic Dermatitis
- Younghee Yun, Hyunho Kim, Wonmo Jung and Bo-Hyeong Jang
•
Session 5-3Dynamic relationship between the emotion and the physical
statues: an observational study through personal health record
- Ye-Seul Lee, Won-Mo Jung, Hyunchul Jang, Sanghyun Kim,
Sun-Yong Chung, Younbyoung Chae
•
Session 5-4From EHR to PHR: a Literature Review
- Ching-Chu Liao, Heng-Shuen Chen, Pei-Ni Chuang,
Yi-Hsuan Huang, Shin-Yu Yeh, Shou-Jen Lan,
Development of Personal Health Records Architecture
for Community Oriented Holistic Care to be in Line
with International Standards
Heng-Shuen Chen1,2, Ching-Chu Liao1, Pei-Shan Yang2,
Yu-Yuan Yu1, Hung-Pin Hou1, Heng-Chang Chen1
1_ Puli Christian Hospital, Puli, Taiwan
2_ National Taiwan University, Taipei, Taiwan
ABSTRACT
In recent years the aging process of Taiwan society become more rapidly. In current health system
referral system was not established and medical
resource was unevenly distributed. Rural and urban
area faced different problems of inaccessibility and
fragmentation of healthcare. Medical expenditure
was escalating due to increasing medical needs of
the elderly and disability.
Objective
ŠŠ
The goals of this study are to explore personal
health records architecture with contemporary
health information technology, and apply the concept of holistic care and community oriented primary healthcare, in order to be in line with the international electronic health records standard.
Methods
ŠŠ
In this study, literature reviews of PHR and EHR
were performed. Renowned experts of EHR and
PHR from foreign countries were invited to visit Taiwan for an international symposium and academic
exchange to share experience. A comprehensive
content architecture was proposed after thorough
discussion and revisions among panels and committee. Implementation of a prototype of PHR system
was be tested in a rural community hospital
most important papers for review. Secondly, Presidents of IMIA (Prof. Hyeoun-Ae Park), AMIA (Prof.
Thomas Payne) and Vice President of JAMI(Prof.
Naoki Nakashima) were invited to attend and give
key note speeches in the international symposium
of consumer health informatics and PHR, Aug 12,
2016, in Puli, Taiwan. In return delegates from Taiwan will visited foreign academic institutions and
attend APAMI 2016 in Korea and AMIA 2016 in US.
Thirdly, a multidimensional architecture of PHR content was proposed with: 1. natural history of the disease and stages of preventive medicine, 2. continum
of healthcare with 6 stages including preventive
care, emergency care, acute care, subacute care,
long term care and terminal care, 3. Holistic care
of bio-psycho-socio-spiritual model of approaches,
4.Assisted living with 6 categories including food,
clothing, housing, transportation, education, and recreation.
Conclusion
ŠŠ
In future, PHR could support not only rural area and
in underdeveloped countries but also inspire the advanced countries troubled with expensive high medical technology and many people lacking health care.
Results
ŠŠ
Firstly, literature searches from academic database
such as PubMed, EMBASE, MEDLINE and document repositories of government and academic
association were retrieved and selected about 50
Keywords: electronic health records, personal health records, holistic care, community medicine
46
2016 Conference of
Asia Pacific Association for Medical Informatics
Development of a Mobile Application in Conjunction with a Web
Service for Personal Health Records regarding Atopic Dermatitis
Younghee Yun1, Hyunho Kim2,
Wonmo Jung3 and Bo-Hyeong Jang4
1_ Department of Dermatology of Korean Medicine, Kyung Hee University
Hospital at Gangdong, Seoul, Korea
2_ Department of Biofunctional Medicine & Diagnostics, Kyung Hee
University Korean Medicine Hospital
3_ Department of Science in Korean Medicine, Graduate School, Kyung
Hee University, Seoul, Korea
4_ Department of Preventive Medicine, College of Korean Medicine,
Kyung Hee University, Seoul, Korea
ABSTRACT
Objective
ŠŠ
Although a myriad of mobile applications that provide health or medical information have been developed for various diseases including atopic dermatitis (AD), most of them are not used by physicians in
clinics. Usually they are used by patients themselves
only to record their symptoms or get information
about diseases. The purpose of this study was to develop a smart system which can record and retrieve
medical symptoms and clinical information generated in daily life in order to connect physicians and patients and to enhance the quality and the efficiency
of medical care.
information related to AD from patients and a web
service to assist physicians for retrieving and integrating symptom information. User interface of mobile application was designed under consideration
of convenient daily input. On web service for physicians, visualization and tabulation of the time-series
recorded symptoms were designed based on the
practical need of physicians. The developed system
has been evaluated from July 2015 to August in
clinical situation.
Results
ŠŠ
The smart phone application, ‘Atopy Mobile Note’,
has been registered in Google Play Store from
March 2016. Patients can record AD score, episodic events of their illness, food diary, treatment
and management diary using their mobile devices.
Recorded information is saved in database of central server immediately. Physicians can retrieve a
sequence of records in the clinic using web service
program.
Conclusion
ŠŠ
Methods
ŠŠ
Task force team consisting of three physicians and
two developers has developed the system. Patient
centered informatics tool for AD was developed by
implementing a mobile application to collect daily
We have developed an mPHR application for patients and corresponding web service for physicians.
To take advantage of this system in health care will
require more research to find out how much it actually more beneficial to patients and/or physicians.
Keywords: Personal Health Records; Electronic Health Records; Atopic Dermatitis, Application
47
Dynamic relationship between the emotion and the physical
statues: an observational study through personal health record
Ye-Seul Lee1, Won-Mo Jung2, Hyunchul Jang3,
Sanghyun Kim4, Sun-Yong Chung5, Younbyoung Chae6
1_ Acupuncture and Meridian Science Research Center, College of Korean Medicine,
Kyung Hee University, Seoul, Korea.
2_ Acupuncture and Meridian Science Research Center, College of Korean Medicine,
Kyung Hee University, Seoul, Korea.
3_ Mibyeong Research Center, Korean Institute of Oriental Medicine, Daejeon, Korea
4_ Mibyeong Research Center, Korean Institute of Oriental Medicine, Daejeon, Korea
5_ Dept. of Neuropsychiatry, College of Korean Medicine, Kyung Hee, Seoul, Korea
6_ Acupuncture and Meridian Science Research Center, College of Korean Medicine,
Kyung Hee University, Seoul, Korea.
ABSTRACT
Objective
ŠŠ
The need to develop a Personal Health Record (PHR)
have increased in order to support the prevention
and management of diseases both inside and outside the medical institutions. Management of chronic disorders on a daily basis is possible by collecting
and analyzing data on the patient’s condition. In this
research, we developed an evaluative form of daily
psychological and physical conditions in the format
of mobile application which we named “Mind Mirror”.
This study aims to show that PHR can be a useful
tool to assess the relationship between emotion and
physical condition.
Methods
ŠŠ
In the development of “Mind Mirror”, parameters to
measure emotional valence and physical state of
pain and fatigue were developed based on different
theories. Using a 30-day observational data from 20
participants using this mobile platform, we analyzed
the relationship between the affective state and
physical condition through cross-correlation analysis of time-series. Based on the cross-correlation
between three parameters, a multilevel multivariate
regression model (MGLM, Multivariate Generalized Linear Mixed Model) with subjects as random
effects was carried out. By selecting model that
minimizes AIC, BIC, and log-likelihood, the dynamics
among three parameters were determined. The relationships between emotion, pain and fatigue was
visualized in a dynamic network structure.
Results
ŠŠ
The strongest cross-correlation in this emotion
and physical state data was at lag 0, which implied that emotion and the body condition changed
concurrently. MGLM model showed that emotion
and fatigue were significantly associated with one
another (emotion → fatigue:β = -0.233, p<0.001;
fatigue → emotion:β = -0.150, p<0.001). Fatigue
and pain were significantly associated one another
as well (fatigue → pain:β =0.250, p<0.001; pain
→ fatigue:β=0.398, p<0.001). Emotion and pain
influenced each other but without significance (emotion → pain:β =-0.021, p=0.063; pain→emotion:β
=-0.022, p=0.063).
Conclusion
ŠŠ
This study implies that the relationship between
emotion and the body interacts instantaneously.
Dynamics between emotion and body through PHR
indicates the need to examine the mind and body as
a whole. This study suggests a possible solution to
manage health through mobile platform on a daily
basis.
Keywords: Personal health record (PHR), time-series analysis, cross-correlation, Multivariate Generalized Linear Mixed Models (MGLM)
48
2016 Conference of
Asia Pacific Association for Medical Informatics
From EHR to PHR: a Literature Review
Ching-Chu Liao1,3, Heng-Shuen Chen1,2, Pei-Ni Chuang2,
Yi-Hsuan Huang2, Shin-Yu Yeh2, Shou-Jen Lan3,
1_ Puli Christian Hospital, Puli, Taiwan)
2_ National Taiwan University Hospital, Taipei, Taiwan)
3_ Asia University,Taichung, Taiwan.
ABSTRACT
Personal health record (PHR) is an innovative health
information technology that empowers patients to
take more active role in their own health and make
informed decisions. PHR systems may be effective
in improving preventive care and plays an important
role in reducing medical costs in this rapidly ageing
Taiwan society.
Objective
ŠŠ
The purposes of this study are to 1.Explore contemporary health information technology and to be
inline with the international electronic health records
for the aging society in Taiwan. 2.Review well established PHR systems and provide a summary of future direction for personal health records in Taiwan.
3.Define the role of family physician in connecting
the patient and the PHR system.
Methods
ŠŠ
Our study was a qualitative literature review and
we tried to obtain full-text and review articles from
PubMed, EMBASE and MEDLINE. Key terms used in
the search included “personal health records,” “PHRs,”
“Electronic personal records,” and “EHRs.”
Results
ŠŠ
Meaningful Use (MU by HITECH Act) emphasized
the three stages EHRs to improve quality, safety,
efficiency, and health inequality. PHRs were in MU
stage 2 patients and their families to be involved in
their healthcare. MU stage 3 is to improve health
outcomes through quality, safety, efficiency, and
access to PHRs and allow communication between
patients and physicians. PHR were applied in UK using Summary Care Records(SCR) and HealthSpace
portal. In the US, PHR in the Veterans Health Ad-
ministration (VHA) system, Google Health by Google and HealthVault by Microsoft were developed.
However, some operations were not so smooth,
and the Google health and HealthSpace were discontinued in 2011 and 2012. The difficulties in
PHR adoption include: 1.Technical barriers, 2. Policy
barriers 3.Computer Competency, Internet Access,
and Health Literacy, 4. Economic and market forces.
For elderly, cognitive and physical impairments in
addition to compromise the situation. Since communication of provider is a important part of patient
portal, there is concern over potential increased
workload. However, it is helpful for family physicians
to directly engage patients by using a portal and
integrate with individual health care. Besides computer, the rise of mobile phone also makes new opportunities to overcome behavior change, cost, legal
concerns, privacy concern, inconvenience, design
shortcomings, and the inability to share information
across organizations. The government should address the issues to ensure the e-health services to
be effectively, efficiently, and safely accessed, to
promote PHRs’ standards, and a series of possible
legislative actions such as tax deduction for whom
adopt PHR.
Conclusion
ŠŠ
The management of a PHR is primarily the responsibility of the patient. To break down the barriers
to PHR adoption the most effective mechanisms
are education and research. Future research on the
field should focus not only in the evaluation of PHR
systems functionality and architecture, but also on
other critical aspects such as accessibility, usability
and effectiveness especially for disabled and elderly
people,
Keywords: personal health records, electronic health records, family physician, review, aging society
49
6
November 3 (Thu), 2016
Oral
Session
Session 6: Clinical Decision Support System
•
Session 6-1Classification of clinical prediction rules by design and usage
- Georgina Kennedy, Blanca Gallego
•
Session 6-2Association of lipid profile & HbA1c among T2DM patients:
A cross-sectional study in Bangladesh
- MM Islam; Tahmina Nasrin Poly; Usman Iqbal; Suleman Atique;
Navneet Kumar Dubey; Phung-Anh Nguyen; Yu-Chuan (Jack) Li;
Shabbir Syed-Abdul
•
Session 6-3The potential use of computer-assisted image processing for
detecting Acute Lymphoblastic Leukemia
- Lutfan Lazuardi, Annisa Ristya Rahmanti, Ika Candradewi,
Sari Rahmawati Kusuma Dewi, Agus Harjoko
•
Session 6-4EMRs- Towards a better User Interphase (UI)
- Dr S B Gogia
•
Session 6-5Impact of a QQ-based Virtual Health Paradise on reducing
cardiovascular disease risk factors among healthy Chinese
adults
- Bin Li, Ying Wu
50
2016 Conference of
Asia Pacific Association for Medical Informatics
Classification of clinical prediction rules by design and usage
Georgina Kennedy, Blanca Gallego
Macquarie University, Australian Institute of
Health Informatics, Sydney, Australia
ABSTRACT
Objective
ŠŠ
With the uptake of Electronic Medical Records worldwide, clinical prediction rules (CPR) have proliferated
in the medical literature, more than doubling between
1995 and 2005. CPRs vary widely in their design
and intended use. This makes them a challenging target when searching the literature. Furthermore, lack
of standards on the evaluation of these tools and the
fact that few have undergone formal impact analysis,
makes it harder for clinicians and researchers to assess their utility and safety in clinical practice.
This work presents a meaningful classification
scheme for CPR research, which may go some way to
crystallising the definition, reducing the inconsistency
of terminology in their publication and pedagogy, and
guiding the generation of appropriate standards for
their evaluation and inclusion in clinical guidelines
Methods
ŠŠ
As part of a more broad systematic review, publications were sought that observed health provider
perspectives on CPR . Descriptions of CPRs as
provided by retrieved publications were grouped by
their methods and intended usage. Comparisons
were made across rules to identify the dimensions
by which CPRs can differ, and to categorise these
differences into meaningful groups.
Results
ŠŠ
CPRs were found to be definable across 3 dimensions – predictor or factor type (inputs), clinical goal
(methods), and intended action (outputs).
I. Predictors: The most significant criterion on
which to assess the input factors is if they are
modifiable. This dictates if the model can be
used actively, to direct the clinician or patient to
change behaviours or treatment in order to target a desirable outcome, or simply to present a
baseline or conditional risk.
II. Clinical Goal: The majority of publications divide
CPRs into diagnostic and prognostic models. A
Keywords: Clinical Prediction Rules
51
small number included in their definition models
which recommended a specific therapeutic path,
however when reviewing the rules themselves,
none were identified which made a treatment
recommendation on their own — more typical is a
guideline-based recommendation that interprets
the rule’s output within a specific context. Therapeutic rules were therefore not considered.
III. Intended Action: The dimension in which there
was the least consensus was that of the intended
action. This is because researchers, policy makers
and clinicians have differences in opinion on how
prediction rules can and should be used in practice. Some clinicians view CPRs as an assistive
tool, useful only for novices and to help patients
understand their status and impact of potential
treatment options. Other stakeholders treat CPRs
as prescriptive, expecting 100% coverage to
meet a specific outcome goal, such as reduction
in unnecessary testing burden.
Conclusion
ŠŠ
The classification of CPRs across the dimensions
of predictor type, clinical goal and intended action is
expected to simplify the literature and solve some of
the issues observed with their search and synthesis.
Figure 1 Proposed classification for CPRs
Association of lipid profile & HbA1c among T2DM patients: A
cross-sectional study in Bangladesh
1,2
MM Islam 1; 4Tahmina Nasrin Poly 2; 1,2Usman Iqbal 3; 1,2Suleman
Atique 4; 5Navneet Kumar Dubey 5; 1,2Phung-Anh Nguyen 6;
1, 2,3
Yu-Chuan (Jack) Li 7; 1,2Shabbir Syed-Abdul 8
1_ Graduate institution of Biomedical Informatics, College of Medical Science
and Technology, Taipei Medical University, Taipei, Taiwan
2_ International Center for Health Information Technology (ICHIT), Taipei Medical
University, Taipei, Taiwan
3_ Chair, Dept. of Dermatology, Wan Fang Hospital, Taipei, Taiwan
4_ Department of Pharmacy, Stamford University, Dhaka, Bangladesh.
5_ Graduate Institute of Biomedical Materials & Tissue Engineering, Taipei
Medical University, Taiwan
ABSTRACT
Objective
ŠŠ
Uncontrolled hyperglycemia due to impaired lipid metabolism has been entanglement with cardiac disease
in diabetic patients. We aimed to explore association
between lipid profile and HbA1c, considering that early detection of lipid, and HbA1c abnormalities could
reduce the risk of cardiovascular disorder in type-2
diabetic’s patients.
Methods
ŠŠ
This cross sectional study included type 2 diabetes
(T2DM) patients who undergone treatment at primary
level hospital in Bangladesh. The study was undertaken
since August 2015 to January 2016 in the inpatient
and outpatient department in three diabetes hospitals
in Dhaka, and 300 subject with T2DM (more than 6
month’s duration) were randomly selected. The patients
with kidney and metabolic bone disorders were excluded
as these conditions might affect the lipid metabolism
in diabetes. We also asked some questionnaire and
checked medical record of each subject regarding age,
body mass index (BMI, Kg/m2), duration of diabetes,
total cholesterol level, low-density lipoprotein (LDL),
high-density lipoprotein (HDL), triglyceride (TG), creatinine level etc. Record were carefully reviewed and
further checked and corrected in case of missing. We
calculated means and standard deviation for continuous
data and percentage for categorical data. Gender-based
patients characteristics were compared using unpaired
t-test and chi-square test for categorical variables. Patients characteristic according to HbA1c quartile were
compared using one-way analysis of variance (ANOVA)
and descriptive analysis for percentage of categorical
variables. The correlation (Pearson’s correlation) studies
were performed between lipid profile and HbA1c. Significance level in each case was set P<0.05.
Result
ŠŠ
Overall, 300 type 2 diabetic patients were enrolled,
out which 89 were males (29.67%) and 211 females
(70.33%). The average age of male and female patients was were 54.65±12.04 and 53.08±10.86
respectively. The mean age of diabetes diagnosis for
male patients was 46.88±10.99 and female patients
was 45.51±10.04. The mean BMI of male and female
patients were 24.66±3.67 and 26.52±4.17 respectively. Among males, the mean value of TC, LDL, HDL,
and TG were 201.42±55.26 mg/dl, 113.30±46.05
mg/dl, 44.66±14.90 mg/dl, 193.48±102.59 mg/
dl respectively while among females, the mean value
of TC, LDL, HDL, and TG were 205.16±55.17 mg/
dl, 115.23±43.91 mg/dl, 45.12±14.03 mg/dl, and
209.98±120.74 mg/dl respectively. A positive relation was observed for male between increasing the
values of HbA1c and BMI, total cholesterol. But in
the case of females, a positive relation was observed
in total cholesterol and creatinine level. Further, in the
females, higher level of HbA1c was also associated
with the greater percentage of smokers, and patients
undergone both oral and insulin medication. There
was significantly positive correlation between HbA1c
and TC (r=.131). Both LDL (r=.025) and HDL (r=.016)
had positive correlation with HbA1c and negative
correlation with TG (r= -.062). However, there was a
linear relationship between HbA1c and dyslipidemia.
Conclusion
ŠŠ
This study showed lipid abnormalities among T2DM
patients with dyslipidemia, hypercholesterolemia
and hypertriglyceridemia. Our study suggest that
early diagnosis of dyslipidemia would be useful to
minimize the unfavorable cardiac risk of type 2 diabetic patients in Bangladesh.
Keywords: Lipid profile, HbA1c, Type 2 diabetes, Dyslipidemia, Hyperglycemia etc.
52
2016 Conference of
Asia Pacific Association for Medical Informatics
The potential use of computer-assisted image processing for
detecting Acute Lymphoblastic Leukemia
Lutfan Lazuardi1, Annisa Ristya Rahmanti1, Ika Candradewi2,
Sari Rahmawati Kusuma Dewi1, Agus Harjoko2
1_ Department of Public Health, Faculty of Medicine, Universitas Gadjah Mada,
Yogyakarta, Indonesia
2_ Department of Computer Science and Electronics, Faculty of Mathematics
and Natural Sciences, Universitas Gadjah Mada, Yogyakarta, Indonesia
ABSTRACT
Acute Lymphoblastic leukemia (ALL) is a cancer that
affects young children and adult. It is characterized
by the proliferation of abnormal number of malignant
and immature white blood cells that can spread in
the bloodstream and other vital organs. In Indonesia
the cases is increasing with relatively high fatality
rate. Early identification of acute leukemia is very
important in reducing the high mortality rate. Currently, acute leukemia detection is conducted using
microscope observation by identifying the morphology and number of blast cells in the peripheral blood
smears cells and bone marrow aspiration. However,
this manual method is time consuming and requires
highly skilled health professionals. This study aimed
to develop an automatic acute leukemia detection
based on digital image processing and pattern recognition. We used Haar classifier and segmentation
based on HSV color space and watershed method
to separate white blood cell and red blood cell in
the peripheral blood smear image. The result using
hough circle transform involving 128 white blood
cell images of acute leukemia patients have produced 78% accuracy.
Objectives
ŠŠ
The study aimed to develop an automatic acute leukemia detection based on digital image processing
and pattern recognition
Methods
ŠŠ
An action research approach introduced by Davison
et al. (2004) was used to develop the system. This
approach consists of several stages i.e: 1) Initiation
phase, to assess the current method for acute leukemia detection and to observe the need to develop
53
a low cost automated early detection 2) Design &
development phase, to develop the prototype using
proposed algorithm for white blood cell (WBC) classification and to conduct system optimization and
adaptation, 3) Testing & implementation phase, to
perform system validation by identifying the normal
and abnormal blood cell (blast cell). 4) Improvement
and Evaluation phase, to evaluate the performance
of the system. The system was developed using
Microsoft Visual Studio 2012 and EMGUCV image
processing library (.NET wrapper of OpenCV) and
C#.NET programming language. Windows Presentation Foundation (WPF) was used as the Graphical
User Interface (GUI) framework of the application.
Digital image processing and pattern recognition
technique were performed to provide information on
the white blood cells or leukocytes morphology from
blood microscopic image. The identification and
classification of white blood cells were divided into
several classical sequences/ stages i.e: 1) Image
acquisition 2) Preprocessing, to improve the quality
of the captured image, 3) Segmentation, to mark
the white blood cells objects as the foreground and
eliminate the background objects (plasma, red blood
cells and platelets), 4) Feature extraction, to obtain
the numerical characteristic of each image characteristic parameter that can be used to detect Acute
Lymphoblastic Leukemia (ALL) or normal blood cell
(shape features and color features), 5) Classification
(pattern analysis) using k-nearest neighbors algorithm, to classify the cells as blast cell or normal
cells. This study has obtained the ethical clearance
from the ethical committee at Sardjito Hospital/Faculty of Medicine-Universitas Gadjah Mada
Results
ŠŠ
To validate the system, we evaluated 128 blood
smear images from already diagnosed ALL patients.
All samples were obtained from Clinical Patology
Laboratory at Sardjito Hospital/Faculty of Medicine-Universitas Gadjah Mada. All blood samples
were digitized under digital microscope Olympus BX41 with 100x magnification and were captured using Optilab Advance and saved in JPG format of size
1280 x 960 pixels. The purpose of segmentation
process is to detect the white blood cells and differentiate with other blood cells. Image segmentation
of white blood cells by using hough circle transform
was done in 128 samples and the system has been
able to detect 100 of the samples and failed to
identify the 28 samples. Based on this, the accuracy
of segmentation process is 78%. The example of
the segmentation result is shown on Figure 1. This
step is also able to identify that white blood cells
roundness is usually more than 75. Another steps
after segmentation process is feature extraction,
which found to be useful in quantifying the parameter of each blood cell type. In general, the shape,
size, roundness and the value of red, green and blue
color of each cell types i.e. neutrophil eosinophil basophil lymphocyte monocyte and lymphoblast are
measured. On the classification stage, k-nearest
neighbors algorithm was implemented and shown
the accuracy of 64% based on the confusion matrix
analysis.
Conclusion
ŠŠ
Our preliminary result showed that the system has
been able to detect the abnormal white blood cells
and differentiate with other normal blood cells. This
feature is potentially useful to help the diagnosis of
patient with Acute Lymphoblastic Leukemia (ALL)
especially in the primary health centers, which usually have limited resources and lack of hematologist
expert. This system may help health practitioners
at primary health centers to identify the abnormal
blood cells and to captured and documented the
blood samples for further consultation with hematologist at referral hospital. Further development and
evaluation is needed to improve the function and to
test the sensitivity, specificity, accuracy and feasibility for application in the clinical setting.
Figure 1 Example of the successful segmentation process
Original image
Detection Process Result of Segmentation
Keywords: Acute Lymphoblastic Leukemia, computer assisted image processing
54
2016 Conference of
Asia Pacific Association for Medical Informatics
EMRs- Towards a better User Interphase (UI)
Dr S B Gogia
President APAMI, President SATHI
ABSTRACT
Background
ŠŠ
Even while Moore’s law and fast connectivity has
ensured faster processing and easy access, data entry remains the slowest link in value chain inhibiting
growth of EMRs and its attendant benefits related
to Meaningful use. Record keeping should be a pleasure and not a pain.
Cost and labour of conforming to meaningful use
requirements has more than offset the incentives
being provided. EMRs are less used since data entry
takes time out from genuine patient care. We have
found that easing entries makes clinicians desirous
to use an EMR as against manual entry or even dictation.
The key is a faster and easier User Interphase(UI),
which can cater to a wide range of individualized
clinician’s needs. This gets eased whenever the
clinicians themselves, like the current author, were
involved in making the EMR. While making the complete solution is beyond the scope of most clinicians,
easy customization conforming to individual needs
is what can and should be provided.
Objectives
ŠŠ
Demonstrate a working solution wherein data entry
is speeded up. Specialized Rheumatology as well as
Plastic Surgery solutions are offered as an example.
Methods
ŠŠ
1. Identify clinician’s needs for a customized solution
2. User developed forms which includes terms they
are familiar with
3. Easing Data reporting and Analysis through
merging with Standards
We had initially made a generalized template for a
few specialities and later expanded the scope to be
configured to a more diseases. Once the user has
identified the specific ICD10, a descriptive template
conforming to the same crops up.
55
This is done through
-- A generic content table with facility for term and
attribute value pairs.
-- - Users create the front end field names matching the terms for which attribute is required
-- - Code compatible attribute lists are created by
the user or preferably selected from SNOMED
CT terms
-- - During data entry time, there is a choice of either of entry through single select or multi select
drop down lists
-- - Once selections are complete, the name value
pairs are placed as together as readable to text.
These are searchable through simple SQL
-- - The system displays only user understandable
terms.
-- - Use of Numeric codes – like those used in
SNOMED we find are not preferred by clinicians
at least in printed reports. If required they are at
available at the backend.
Results
ŠŠ
In addition to minimal keystroke access and detault
entries, EMR adoption has been eased by vendor assisted but user created templates.
Software use has resulted in a jump in patient turnover patient satisfaction as well as better research
output. The methods developed for making data entry easier faster and easier has been appreciated.
Such forms have helped us to widen the scope of
our existing user base to newer specialities.
Conclusion
ŠŠ
A software that is not used can be deemed useless.
Clinicians need to be involved in EMR creation for
better adoption.
Impact of a QQ-based Virtual Health Paradise on reducing cardiovascular disease risk factors among healthy Chinese adults
Bin Li, Ying Wu
School of Nursing, Capital Medical University,
Beijing, China
ABSTRACT
Objective
ŠŠ
Peer education and support have become increasingly attractive as intervention strategies to combat
cardiovascular disease (CVD) risk factors. Considering communication via Instant Messaging (IM) tools
has become a main media for people to socialize,
using QQ, a popular IM platform in China, we established ‘Virtual Health Paradise’ (VHP), an internet communication group to facilitate information
exchange on healthy lifestyle. This study aimed to
evaluate the effectiveness of VHP in the reduction
of CVD risk factors.
Methods
ŠŠ
Weekly face-to-face educational sessions regarding CVD and healthy lifestyle were offered to employees of a Chinese company in Beijing, China
for four weeks. Based on two different work sites,
employees who attended all educational sessions
and possessed at least one CVD risk factor were
blindly assigned into intervention group (IG) and
control group (CG). After collection of baseline data
(including demographics, lifestyle, perceived stress,
CVD related physical examination and blood tests),
interactive and personalized counseling on lifestyle
modification and CVD risk factors was provided to
IG via VHP. Utilizing personal experiences and strategies to establish and sustain healthy lifestyle were
frequently shared within IG for 5 months, while CG
had no access to VHP. Framingham risk score was
introduced as the primary endpoint. The second set
of data was collected at the end of the VHP exposure period.
Results
ŠŠ
The mean age of 156 participants was 34±5.58
(Male: 63.5%, Female: 36.5%). All baseline data
were comparable between CG (n=76) and IC (n=80).
At the end of the study, the reduction of Framingham risk score in IG (-2.43) was significantly larger
than in CG (-0.42, P <0.001). (Figure 1.) The chisquare test revealed that the abnormal rates of BMI,
triglyceride, cholesterol and blood sugar in IG were
significantly lower than those in CG (p<0.05). IG had
significantly higher rates of having a more healthy
diet and moderate exercise than CG (P<0.001).
Contrary, no significant difference was found in CG
when pre-post comparison was made.
Conclusion
ŠŠ
QQ-based VHP could promote a healthy lifestyle
and reduce CVD related risk factors.
Figure 1 P
re- and Post- intervention comparison of CVD
risk factors in IG
Note: AR: abnormal rate, BMI: Body mass index, TC: total cholesterol, TG: triglycerides, FBG: fasting blood glucose
* p<0.05
Keywords: cardiovascular disease; Virtual Health Paradise; lifestyle; risk factors
56
November 4 (Fri), 2016
Oral
Session
7
Session 7: E
lectronic Health Record (EHR)
57
•
Session 7-1The Overview of Free/Libre and Open Source Software in
Medical Domain, from Asia and the World.
- Shinji Kobayashi, Seong K. Mun
•
Session 7-2Strengthen cancer surveillance in Sri Lanka by implementing
Cancer Registry Informatics to Enhance Cancer Registry Data
Accuracy, Completeness, and Timeliness.
- Kamal Wasantha Kumara Seneviratne
•
Session 7-3Determining Morbidity Pattern First Step Deal with Big Data via
Electronic Health Record of Primary Healthcare in Indonesia
- Dina Nur Anggraini Ningrum, Yu-Hsuan Shao, Chien-Yeh Hsu,
Suleman Atique, Hanif Pandu Suhito, Yu-Chuan (Jack) Li
•
Session 7-4An Viewing System for Health Check-up Result Report using
FHIR, CDA IG, and MHD
- Kim Min Gyu, Kim Dae Young, Il Kon Kim
The Overview of Free/Libre and Open Source Software
in Medical Domain, from Asia and the World.
Shinji Kobayashia, Seong K. Munb
a_ The EHR Research Unit, Kyoto University, Kyoto, Japan
b_ Open Source Electronic Health Record Alliance (OSEHRA),
Virginia, USA
ABSTRACT
Free, Libre and open Source Software(FLOSS) has
been one of the key player of ICT development in
the Internet era. Although, most of the OSSs have
started form product by volunteer community,
some FLOSS products have been developed by
commercial companies for commercial uses.
In medical domain, FLOSS products were widely adopted and anticipated that have effects to
reduce high cost of medical information system
and to avoid ‘vendor-lock-in’. In Asia, Japan medical association has promoted their information
infrastructure as FLOSS in the ORCA project from
2000, and AeHIN(ASEAN eHealth Information Network) declared to use FLOSS, DHIS2 and OpenHIE
as their working base.
Nevertheless, there are much problem on the medical use of FLOSS, such as the sustainable business models. In this session, we review the current
status of FLOSS in medical domain in Asia and the
world, and discuss the future work on FLOSS for
our benefit.
Keywords: FLOSS, Open Source Software,
58
2016 Conference of
Asia Pacific Association for Medical Informatics
Strengthen cancer surveillance in Sri Lanka by implementing
Cancer Registry Informatics to Enhance Cancer Registry Data
Accuracy, Completeness, and Timeliness.
Kamal Wasantha Kumara Seneviratne
Ministry of Health, Nutrition & Indigenous Medicine, National
Cancer Control Programme, Colombo, Sri Lanka
ABSTRACT
Objective
ŠŠ
A total of 13,635 new cancer cases had been diagnosed in 2007 with a crude cancer incidence rate
of (CR) 68.0 per 100,000 population. There were
16,888 new cases diagnosed in 2009 with a CR of
82.6. It is evident that the cancer prevalence in Sri
Lanka is on the rise. The National Cancer Control
Programme (NCCP) of Sri Lanka was established in
1980 based on the recommendation made by WHO
after a detailed study on mortality and morbidity
of cancers in Sri Lanka. NCCP is the national focal
point for prevention and control of cancers in the
country. Sri Lanka Cancer Registry (SLCR) plays a
pivotal role in cancer control and is maintained by
NCCP. The value of a cancer registry depends on the
quality of its data. Cancer control planning without
high quality cancer registry data from the cancer
registry leads to misplaced emphasis and wasting of
investment.
Since a considerable number of private health care
institutions and government hospitals have started diagnosis and treatment of cancer patients, the
necessity of using information technology in cancer
surveillance domain to solve the problems for collecting, storing, processing, and analysing of cancer
registry information received from those institutions
have become absolutely essential in order to improve the data quality, accuracy and completeness.
Methods
ŠŠ
Cancer surveillance informatics is the systematic
application of information and computer science
and technology to cancer surveillance practices,
research, and learning. It deals with the resources,
devices, and methods required to optimize the acquisition, storage, retrieval and use of cancer sur-
veillance information for above mentioned purposes.
Thus, using emerging technology to incorporate
automated process and electronic data exchange
in cancer surveillance business is an efficient, fast
and cost effective way to obtain quality, accurate
and complete cancer registry data. Therefore the
new web-based application replaces the current paper-based method which covers about 80% of data
collection and a standalone database system which
covers remaining 20% of data collection. The basic
software infrastructure will be based on Free and
Open Source Software with the two-tier client-server system architecture.
Results
ŠŠ
The system is able to capture real time cancer registry data from 09 provincial Oncology and Surgical
Oncology units, 67 Histopathology and Haematology laboratories and 25 Oral and maxillofacial units.
It will increases the efficacy in collecting, storing,
processing and analyzing of data, improving data
quality, accessibility, timeliness and completeness,
reducing data redundancy, time, and resource consumption.
Conclusion
ŠŠ
The system will ensure timely availability of cancer
registry information that is needed for better improvement in cancer surveillance. Among the goals
expected to achieve are: determine the incidence of
cancers with respect to geographic and demographic characteristics, monitor trends and pattern of
cancer incidence over time, identify high risk populations, provide data for epidemiological studies, and
prioritize health resource allocations.
Keywords: National Cancer Control programme (NCCP); Cancer Incidence; Sri Lanka Cancer Registry; Cancer registry data; Cancer Surveillance Informatics
59
Determining Morbidity Pattern First Step Deal with Big Data via
Electronic Health Record of Primary Healthcare in Indonesia
Dina Nur Anggraini Ningrum1, 2, Yu-Hsuan
Shao1, Chien-Yeh Hsu3, 4, Suleman Atique1,
Hanif Pandu Suhito5, Yu-Chuan (Jack) Li1, 6
1_ Graduate Institute of Biomedical Informatics, College of Medical
Science and Technology, Taipei Medical University, Taipei City, Taiwan
2_ Public Health Department, Semarang State University, Semarang City,
Indonesia
3_ Department of Information Management, National Taipei University of
Nursing and Health Sciences, Taipei City, Taiwan
4_ Master Program in Global Health and Development, Taipei Medical
University, Taipei City, Taiwan
5_ Semarang City Health Department, Semarang City, Indonesia
6_ Department of Dermatology, Wan Fang Hospital, Taipei City, Taiwan
ABSTRACT
Semarang City is a miniature of Indonesia with a
population of 1.5 million citizens. Electronic Health
Record (EHR) was implemented in 37 public primary healthcare centers (PHC) in 2004 with limited
resources and new challenges. One of the major
challenges after 12 years’ implementation of EHR is
to analyze the morbidity pattern with big scale and
velocity data.
Objectives
ŠŠ
We determining the morbidity pattern by age and
gender in the Semarang City via EHR data.
Methods
ŠŠ
We used EHR data of 37 PHC’s from 2011-2014
comprising around 1.8 million visit records. After
validation we got 1,632,782 valid visit records. We
did descriptive analysis of prevalence rates based
on ICD-10 (21 chapters) in each of 11 age group
i.e. baby, child under 5 years old, 5-14, 15-24, …,
till >85 years old and gender (male and female). We
used Spearman Rank Correlation to determine the
correlation between the prevalence rates of each
ICD-10 chapter with the age group. We used Mann
Whitney U Test to compare the prevalence rates of
each ICD-10 chapter by gender. All of the analysis
generated using SAS.
old. The morbidity pattern has been shown in Figure
1. Diseases of the respiratory system are dominated
in age range of 0-54 years old. Above the age of 54
years old, diseases of circulatory system are dominant. Diseases related with injury and external causes have presented significant differences between
both gender. There are diseases groups which are
not correlated with the age (diseases related with
blood, mental and behavioral, ear, digestive system,
pregnancy, injuries, and external causes). Diseases
related with circulatory system, endocrine and metabolism, eye, musculoskeletal, neoplasms, nervous
system, and genitourinary system have shown positive correlations with the age. Diseases related with
skin, symptoms and findings, perinatal period, health
status, infectious, respiratory system, and congenital malformations have been found to be negatively
correlated with the age.
Conclusion
ŠŠ
From this morbidity pattern we know which diseases
groups are influenced by the gender and age. This information can be used in PHC as tool to predict the
morbidity pattern and to help public health decision
making based on gender and age to make future intervention and strategies.
Results
ŠŠ
The visit by patients was dominated by female
(63.78 %) overall, especially in age 15 - 74 years
Keywords: Public health informatics, morbidity pattern, primary health care, electronic health care, ICD-10
60
2016 Conference of
Asia Pacific Association for Medical Informatics
Figure 1 Morbidity Pattern in Primary Healthcare Semarang City, Indonesia 2011-2014
Table1 S
pearman Rank Correlation for Prevalence Rate of 21 ICD-10 Cluster and 11 Age Groups
Keywords: Public health informatics, morbidity pattern, primary health care, electronic health care, ICD-10
61
An Viewing System for Health Check-up Result Report using
FHIR, CDA IG, and MHD
Kim Min Gyu, Kim Dae
Young, Il Kon Kim
School of Computer Science & Engineering, Kyungpook
National University, Daegu, Republic of Korea
ABSTRACT
Objective
ŠŠ
Individual’s demand for self-health care is increasing
with the advance of health care services. To meet
the request, the Ministry of Health and Welfare, and
National Health Insurance Service, Korea provides
regular health check-up service for all the people
over 40 years old. There are some major problems:
health check-up result report by snail mail, transfer
of the check-up result report to the healthcare service provider by patient themselves, no interoperability because of applications not compliant with
standards. To solve these problems, we developed
the mobile health service based on international
health information exchange standards which helps
the subject receive and view the check-up result report in both mobile and desktop environment.
Methods
ŠŠ
The system consists of three components. Each
component complies with each domain’s standards:
HL7 CDA(Clinical Document Architecture) IG(Implementation Guideline) based on C-CDA(Consolidated-CDA) for Check-up Result, FHIR(Fast Health
Interoperability Resources) for health data transfer
using MHD(Mobile access to Health Document profile) and HTML5, CSS3 for CDA Viewing. The CDA
IG supports up to entry level and refers to LOINC(Logical Observation Identifiers Names and Codes)
for terminology standard. We use the CDA IG generation library developed in a previous our study. We
adopted the Spring Framework for this mobile health
server. The Server transfers this health check-up
result report CDA to viewer application using RESTful services by MHD transactions and uses Bundle,
DocumentManifest, DocumentReference resources
specified by FHIR. The viewer is based on checkup result report format which the Ministry of Health
and Welfare provides. The viewer displays this health
check-up result report using XSL with HTML5 and
Google chart library.
Results
ŠŠ
The system may be operated in a series of processes such as clinical document generation, transmission and viewing. But for now, the service is only
applied to the 1st general health check-up result report. This check-up CDA generator library and viewer are in testing and stabilization phase in Android
Platform. This interoperable data in Health check-up
result report is very valuable for the health big data
analytics services. This system is being applied to
the real data in HESEL related some hospitals.
Conclusion
ŠŠ
We used the international standards for the system
implementation such as CDA, FHIR, HTML5, etc.
The system guarantees the interoperability of contents using standard document format and is platform independent. This system can serve as a data
pre-processing for analyzing health data which is
produced by standardized and structured data format. Using the same interface of standards-based
data transmission method like FHIR, it is very useful
in health care application ecosystem establishment.
We will upgrade the system to support the 2nd general health check-up result report. The architecture
of the system can be applied to other health care
systems related to clinical data exchange. We are
considering to applying not only national health
check-up but also comprehensive health check-up
done by hospitals.
Keywords: Health Check-up, HL7 CDA, HL7 FHIR, MHD Profile, Data Exchange
62
8
November 4 (Fri), 2016
Oral
Session
Session 8: m
Health for Connected Patient Care
63
•
Session 8-1Improving Patients’ Medication Adherence and Lipid Level by
Customized Short Message: a Six-Month Clinical Trial
- Yisi Liu, Yang Gong, Yiran Li, Qiannan Zhang, Boya Yu,
and Ying Wu,
•
Session 8-2Impact on Effectiveness of Chronic Disease Management in a Rural
Hospital of middle Taiwan by a Case Management Information
System
- Mei-Chih Chen, Heng-Shuen Chen, Ching-Chu Liao, Chin-Yao Chen,
Ting-Hsuan Chen , Pei-Chia Lo, Hung-Pin Hou, Yu-yuan Yu
•
Session 8-3Remote Monitoring in Home Peritoneal Dialysis:
Seamless User-centered Proactive Provision Of Risk-stratified
Treatment in Peritoneal Dialysis (SUPPORT-PD)
- Oommen J, Kazem R and Vivekanand J
•
Session 8-4Disruptive Innovation in mHealth: A case of Nutrition Monitoring
from a LMIC
- Pamod M. Amarakoon and Roshan Hewapathirana
Improving Patients’ Medication Adherence and Lipid Level by
Customized Short Message: a Six-Month Clinical Trial
Yisi Liua, Yang Gongb, Yiran Lia,
Qiannan Zhanga, Boya Yua, and Ying Wua,
a_ School of Nursing, Capital Medical University, Beijing, China
b_ The University of Texas Health Science Center, Houston, USA
ABSTRACT
Objective
ŠŠ
To explore the effect of customized short message
service (SMS) over six months on medication adherence, lipid level, and clinical outcome of patients with
acute coronary syndrome (ACS) after percutaneous
coronary intervention (PCI).
Methods
ŠŠ
In this clinical trial, patients with ACS after PCI were
randomly assigned into an intervention group and
a control group. A health educational kit, based on
ACS secondary prevention guideline, was designed
with five themes, i.e. medication, lipid, exercise, diet,
smoking and drinking. A SMS pool was generated on
the basis of the health education kit and customized
for each study subject according to his/her geographic data and health history. Participants of the
intervention group received one individualized SMS
per week, 24 SMS in total over six months. No SMS
was sent to the control group. Primary outcomes
were measured at the baseline and sixth month, including medication adherence measured by Morisky
Scale, lipid level by intravenous blood sample, readmission rate, and cardiovascular events, comprised
of the occurrence of unstable angina, ST-segment
evaluation myocardial infarction, non-ST-segment
evaluation myocardial infarction, and coronary revascularization.
Results
ŠŠ
68 patients completed the entire follow-ups, with
34 in each group. No significant difference was
observed in baseline geographic and clinical characteristics (age, gender, education, medical history,
etc) between two groups. After six-month SMS
intervention against the control group, the intervention group exhibited higher medication compliance
(59.4% vs 90.60%, P=0.004), borderline lower
triglyceride level (1.67±1.04 vs 1.27±0.50 mmol/l,
P=0.061), and a downward trend in the occurrence
of CV events (40.60% vs 21.90%, P=0.106). Prepost test showed low density lipoprotein cholesterol
(LDL) level was significantly lower than the baseline
in the intervention group (2.02±0.67 vs 1.83±0.56,
P=0.027), while no difference was observed for the
control group. See Figure 1.
Conclusion
ŠŠ
Continuous customized SMS could help patients
with ACS after PCI improve medication compliance
and better manage their dyslipidemia. Smart SMS
platform with customized SMS appears valuable in
CVD secondary prevention. Effects longer than 6
months are worth further investigations.
Figure 1 Effects of six months SMS on medication adherence, TG, and LDL. TG=triglyceride; LDL=low density lipoprotein cholesterol
Keywords: Short Message Service; Medication Adherence; Cholesterol; Acute Coronary Syndrome
64
2016 Conference of
Asia Pacific Association for Medical Informatics
Impact on Effectiveness of Chronic Disease Management
in a Rural Hospital of middle Taiwan by a Case Management
Information System
Mei-Chih Chen1, Heng-Shuen Chen1,2, Ching-Chu Liao1, Chin-Yao Chen1,
Ting-Hsuan Chen1 , Pei-Chia Lo1, Hung-Pin Hou1, Yu-yuan Yu1
1. Puli Christian Hospital, Puli, Taiwan
2. National Taiwan University Hospital, Taipei, Taiwan
ABSTRACT
Morbidity compression hypotheses burden of lifetime illness may be compressed into a shorter period before the time of death. So healthcare should
be started at a earlier period in elderly and chronic
diseases. In this study a case management information system for chronic diseases was developed in
Puli Christian Hospital using cloud computing technology to monitor checking points and quality index,
and provide automated reminders for instant medical management.
Objectives
ŠŠ
The purposes of this study are to 1. understand the
impact on time efficiency of chronic disease management by case manager, and 2. understand the
impact on chronic disease complication (diabetes
mellitus as an example) after introduction of the
case management information system.
Methods
ŠŠ
First, time duration of chronic disease case management of each patient and the total patient visits
in 2014 before introduction of case management
information system and after in 2015 were measured. Second, diabetic patients received medical
care in Puli Christian hospital from 2014 to 2015
were retrieved and divided to managed group included in case management information system and
non-managed group which was not included. Statistic analysis with SPSS20 were performed including
complete rate of fundus examination and percentages of diabetic complications including retinopathy,
peripheral neuropathy, cataracts, peripheral vasculopathy and nephropathy.
Results
ŠŠ
In 2014 the number of patient visits to a case manager was1,779 . In 2015 after system launched,
the number of patient visits became 1,900 with increase of 121, and the average duration of visit was
2.4 minutes less than in 2014. 73 hours was saved
per case manage after using information system.
In addition to the time efficiency, case managers
could expand their service from one disease to two
diseases. With integration and transparency of the
system, case manager can understand the overall
health condition of the patients and share health
information with other medical professionals to
achieve effective communication and holistic health
care. From 2014 to 2015, 1,764 diabetic patients
received regular medical care in Puli Christian Hospital divided to managed group 886 and non-managed
group 878. The average age was 63.34 and 51.8%
of them are male. Statistics analysis revealed relative higher retinopathy 16% (P <0.001) in managed
group, however the complete rate of fundus examination 28.9% (P <0.001)was even higher..
Conclusion
ŠŠ
Chronic disease patients in rural areas would have
more complications due to relative poor access to
medical resources. Case management is important
for morbidity compression in chronic diseases, however time consuming and complicated care flow for
multiple diseases management are burdens for case
managers. Introduction of case management information system can not only reduce the service time
for each patient visit but also enhance the complete
rate of examinations for disease complications by
automatic reminders especially useful in rural area.
Keywords: chronic disease, case management information system, reminder, effectiveness
65
Remote Monitoring in Home Peritoneal Dialysis: Seamless
User-centered Proactive Provision Of Risk-stratified Treatment
in Peritoneal Dialysis (SUPPORT-PD)
Oommen J, Kazem R and Vivekanand J
1_ The George Institute for Global Health India, New Delhi, India.
2_ The George Institute for Global Health UK, University of Oxford, UK
3_ The George Institute for Global Health India & James Martin
Fellow, Oxford, UK
ABSTRACT
Objective
ŠŠ
To design a remote health monitoring and non-pharmacological self-management system that has
good usability among home based peritoneal dialysis patients using a mobile application. Approximately 200,000 Indians develop end stage renal
disease each year. As in many parts of the world,
hospital-based hemodialysis is still the main modality of treatment for patients with ESRD. Reduced
frequency of contact with healthcare providers is
considered to be a barrier to more widespread use
of home-based dialysis therapies. Home monitoring
using smartphone enabled health solutions will allow
early recognition of problems, reduce morbidity and
improve treatment adherence in patients on peritoneal dialysis (PD).
Methods
ŠŠ
Low and high fidelity qualitative usability feedback
was collected from PD patients by semi-structured
interviews. User needs and perspectives on PD
were transcribed into software requirement specifications, which was used for the solution design.
Prototype application screens were built using Java
eclipse and deployed on 10-inch Android tablets,
respondents asked to evaluate the functionality and
features, with iterative refinement of features. The
mobile interface was build an android native application with support for android 5.1, SQLite database
and AnD APIs for interfacing with external blood
pressure monitors. Web services were built using
the bootstrap framework, with PHP and MySQL for
the server side. The application has been built to ensure intense engagement by the patients and highly
interactive step by step decision support to ensure
compliance to treatment guidelines and feedback
in case of any complications or deviations in the parameters such as weight / BP / Fluid balance.
Results
ŠŠ
We present the end user defined design approaches, technological design and preliminary evaluation
of the SUPPORT PD application. A total of 10 patients (mean age 60 years, time on PD 2.5 years) on
peritoneal dialysis at two centers in India (Delhi and
Hyderabad) were engaged in the usability reviews.
All respondents were satisfied with PD, reasonably
independent and managing their therapy. Current
practice included record of fluid intake, output and
number of exchanges. Patients appreciated the value of a remote monitoring tool and described the
desirable features (exchange details, detect complications, guidance of changes in therapy, ease of input, ability to review progress on certain parameters,
alerts for impending complications, automatic information to treating team and educational materials).
About 60% responded that they would use the device themselves, others would be supported by the
immediate caregiver. The study is currently ongoing
and recruiting subjects for home based monitoring
while on PD.
Conclusion
ŠŠ
Patient centered M-health Interventions are feasible
and considered of value by patients on home-based
dialysis. Engaging with the end-users early in the design process gave us insights on patient preferences
for home based monitoring. We propose to expand
the SUPPORT PD application to other chronic diseases and evaluate usability and user acceptance to
provide a comprehensive continuum of care.
Keywords: Remote health monitoring, m-health, clinical decision support, chronic diseases
66
2016 Conference of
Asia Pacific Association for Medical Informatics
Disruptive Innovation in mHealth: A case of Nutrition
Monitoring from a LMIC
Pamod M. Amarakoon1 and Roshan Hewapathirana2
1_ Nutrition Coordination Division, Ministry of Health, Sri Lanka
2_ Health Informatics Society of Sri Lanka
ABSTRACT
Objective
ŠŠ
Results
ŠŠ
Methods
ŠŠ
Conclusion
ŠŠ
Malnutrition has been a major public health concern
in Sri Lanka, and all Low and Middle Income Countries (LMIC). Care for malnourished children in Sri
Lanka is provided primarily at field level by Public
Health Midwife (PHM). Paper based collection of
data at field level is currently used to generate nutrition information. This result in major data quality issues, especially related to multi sector collaboration
needed in nutritional interventions. The objective of
the study is to introduce Disruptive Innovation (DI)
to current information flow to improve the process in
long run.
The DI is the process of developing new products
or services to replace existing technologies and
gain a competitive advantage. With the DI principles
in mind, an Android based mobile application was
developed with primary aim of registering malnourished children, routinely following up their nutritional
parameters at field level and assessing risk factors
at households leading to malnutrition. The aim of this
approach to challenge the conventional paper based
data collection despite not having a felt need for a
mHealth solution in State health sector. The mobile
app was implemented in 3 vulnerable districts of Sri
Lanka during the pilot phase of electronic data collection. The implementation was followed by series
of focus group discussions (n=540).
Several major issues of the traditional paper based
system were revealed by PHMs even though this was
not in the agenda of higher officials. The mHealth
approach was not only challenging the paper based
records, but also a paradigm shift from the desktop
based electronic data gathering taking the data
collection process to the field level. PHMs were of
opinion that use of smart phone for data entry was
a value addition for their job role with a better social
recognition, apart from the new functionalities for
real-time nutrition monitoring and improved mobility
compared to desktop applications.
It was evident that the mobile technology could
be used as a DI approach in State health sector of
LMICs despite the resource limitations. mHealth
challenges both the paper based record systems as
well as Desktop based data collection. Even though
the initial cost of mHealth solutions may be higher,
mobile devices have the advantage of a potential DI
candidate for the mobile phone being able to use for
multiple purposes beyond mere data collection.
Keywords: Disruptive Innovation, mHealth, Child Nutrition Monitoring, LMIC
67
9
November 4 (Fri), 2016
Oral
Session
Session 9: ( APAMI Section) Public health &
National Health Strategy
•
Session 9-1History of Health IT in Japan: Invitation to IMIA’s 50th
anniversary
- Michio Kimura
•
Session 9-2Climate & Health Innovation for Substainable Smart Cities
- Adam CHEE
•
Session 9-3The development and implementation of Stroke risk prediction
model in National Health Insurance’s Personal Health record
- Jae-Woo Lee, Hyun-sun Lim, Dong-wook Kim, Kyung-Hee Cho
•
Session 9-4Dialysis Outcomes in India:
An online clinical outcomes registry for dialysis outcomes,
lessons learned and opportunities for scale up
- Oommen J, John K and Vivekanand J
•
Session 9-5State of Management Information System (MIS) in Tertiary Care
Hospitals in Pakistan
- Suleman Atique, Abdallah Ahmed Elbakkoush,
Syed Muhammad Mursalin, Ting-Wu Chuang, Chien-Yeh Hsu,
Shabbir Syed-Abdul
68
2016 Conference of
Asia Pacific Association for Medical Informatics
History of Health IT in Japan
- Invitation to IMIA’s 50th anniversary Michio Kimura, MD, PhD, FACMI
Hamamatsu University, School of Medicine, Hamamatsu, Japan.
ABSTRACT
Objective
ŠŠ
To record Japan’s history of health IT, also for the
purpose of inviting other APAMI members participation to IMIA 50th Anniversary publication and special session at MEDINFO 2017 Xiamen.
Methods
ŠŠ
Tokyo University, Chiba University, Hamamatsu University, and Ministry of health had materials. Prof.
Kaihara summarized the 1994 situation with issues
to be solved.
Results
ŠŠ
IT in healthcare in Japan started in 70’s, as an automation of hospital work. It started at some departments, such as pharmacy, laboratory, and accounting, all of which are by standalone computers.
Tokyo University Hospital started pharmacy claim
messaging in 76, a first case of connections between departments.
In 1980, MEDINFO80 was held in Tokyo, with Prof.
Roukins as IMIA president and Prof. Collen as SPC
chair. In these years, order communication system
was starting to be introduced. Also some trials to
record patient charts were found in the conference
contents.
In 1985, Chiba University Hospital started recording lab chemical results from the order entry system
with some clinical findings. The data can be retrieved
now in 30+ year history chart.
Also, in 80’s, radiology examination images became
available outside of radiology department as PACS.
In 1988, first hetero-vendor connection between
GE CT and Toshiba PACS was implemented conforming ACR-NEMA standard (now DICOM as version 3).
In 90’s order entry system became popular. Then,
Keywords: History, IT in Healthcare, IMIA 50th anniversary,
69
Kameda General Hospital, Kochi Medical University
are the first runners of electronic medical records.
In 1994, Prof. Kaihara summarized the health IT, and
stated “Coming 5 years problems for Medical Informatics”, as follows.
Medical informatics (in Japan) developed along
with hospital information systems. Vendor efforts and technical innovation made, say, high
installation of order entry systems.
We can be proud of this, but only few innovations
in our field came true. HIS has huge clinical data,
but they are not yielding new clinical evidences,
being unified database
with other databases. Nor our system can exchange data between other providers.
Now, survey in 2014, 34% of hospitals (86% in
400+ bed large hospitals), 35% of clinics have
electronic medical records. 865 hospitals have HL7
based standardized patient data storages. It is hard
to find non-DICOM imaging apparatus. 200+ regional EHR are in operation. Reimburse claim data became available for researchers in 2013, many years
behind Korea and Taiwan.
Conclusion
ŠŠ
Following are the issues raised by Prof. Kaihara in
1994, and my ratings now.
• Exchanging needs and seeds among user, vendor,
government (B+)
• Public recognition that “Soft is mightier than the
Hard” (C-)
• Standardized codes, messages, protocols (B+)
• Reimbursement premium on information exchange between providers (B-)
• EMR, instead of papers, films (B+)
• Medical record given back to patients (C+)
Climate & Health Innovation for Substainable Smart Cities
Prof.Dr. Adam CHEE
President, Association for Medical and Bio-Informatics, Singapore (AMBIS)
ABSTRACT
Category
ŠŠ
ЕЕ Public health & National Health Strategy
ЕЕ APAMI Session
Climate change brings forth environmental consequences affecting clean air, safe drinking water, food
supply and secure / safe shelter, all of which are factors affecting (directly and/or indirectly) the physical,
social, and psychological health of humans
Rapid urbanisation with poor urban planning had
without doubt, played a key role in exacerbating the
problem. However, with the advent of Smart Cities,
one can take the opportunity, not only to mitigate
direct damages to health caused by climate change
but also to reduce the damage (such as lowering
emissions of greenhouse gase) by turning “cities as
problems” into “cities wtih as solutions”.
This presentation aims to address how Smart Cities
can adopt effective Public Health Informatics with
Digital Health(care) as part of their health strategy
to ensure health substainability and effective outcomes.
70
2016 Conference of
Asia Pacific Association for Medical Informatics
The development and implementation of Stroke risk prediction
model in National Health Insurance’s Personal Health record
Jae-Woo Lee1,2, Hyun-sun Lim2,
Dong-wook Kim2, Kyung-Hee Cho1
1_ Department of Family Medicine, National Health Insurance Service
Ilsan Hospital, Goyang-si, Republic of Korea
2_Department of Policy Research Affair, National Health Insurance
Service Ilsan Hospital, Goyang-si, Republic of Korea
3_ Department of Family Medicine, National Health Insurance Service
Ilsan Hospital, Goyang-si, Republic of Korea
ABSTRACT
Objective
ŠŠ
The National Health Insurance Service (NHIS) in
Korea which owns the entire national health examination data is providing the Personal Health Record
service called My Health Bank in the health information website - Health iN (http://hi.nhis.or.kr/main.
do). The purpose of this study is to upload the stroke
prediction model based on NHIS data to the disease
prediction program of My Health Bank.
Methods
ŠŠ
After constructing the data set come from national
general medical examinees of 6,885,789 people
(male 3,816,231, female 3,069,558) in 2002 ~
2003, it identified when stroke (ICD-10 criteria)
was first diagnosed among the inpatient and outpatient until 2013, then developed a 10-year predicted probability and estimated the stroke prediction
model using Cox’s proportional hazard model.
Results
ŠŠ
Risk factors in stroke risk model, include the age,
the square of the age, BMI, cholesterol, blood pressure, blood glucose level, smoking status and intensity, physical activity, alcohol intake, past history
(hypertention, coronary heart disease, stroke) and
family history (stroke, coronary heart disease). The
AUC values of stroke risk prediction model from the
Keywords: Stroke risk, Risk prediction model, national health data.
71
external validation data set were 0.827 in men and
0.822 in women, which showed a high predictive
power. The predicted values of Hosmer and Lemeshow’s χ2 test were 186.19 in man and 111.21 in
women showing little difference between observation probability and predicted probability. Using a
risk prediction model of stroke, this study calculated
the national health examination data automatically
read and uploaded it’s results to Health iN website
while developed the algorithm to provide a personalized message on the basis of each user’s data. In
addition, it developed a lifestyle correction message
about the stroke risk factors that required improvements.
Conclusion
ŠŠ
This study developed the stroke risk prediction
program based on the nation medical examination
results. By doing so, it intended to more effectively
give medical users the motivation of health management and to induce changes in their health behaviors.
Dialysis Outcomes in India: An online clinical outcomes registry
for dialysis outcomes, lessons learned and opportunities
for scale up
Oommen J, John K and Vivekanand J
1_ The George Institute for Global Health India, New Delhi, India.
2_ The George Institute for Global Health, Sydney Australia
(On behalf of Study coordinating group)
3_ The George Institute for Global Health India & James Martin
Fellow, Oxford, UK
ABSTRACT
Objective
ŠŠ
There is limited systematic data on clinical outcomes and economic impacts of maintenance dialysis for end stage kidney disease (ESKD) in India and
other emerging economies. To follow up patients
commencing hemodialysis at two North Indian nephrology centers prospectively for 12 months using
a cloud based open source electronic data capture
tool to study the clinical and economic outcomes.
Methods
ŠŠ
A web based electronic data collection and project
monitoring tool was developed using the OpenClinica platform. Baseline demographic and clinical
outcome data were collected comparable to those
recorded by established dialysis registries as well as
data on direct, indirect costs and economic impact
on the patient and family. (include the setting for the
study, the subjects, and the type of statistical analysis; if references are needed, they should be given in
the text)
Results
ŠŠ
Here we present the technological design and implementation of the online data capture tool for
establishment of the dialysis clinical outcomes
registry and results of the 6 month interim analysis.
A total of 119 patients (82 male, 37 female) have
been enrolled thus far, 70 at a public hospital (Chandigarh) and 49 at a private hospital (Delhi). Median
age at enrollment was lower at the public hospital
compared to the private (37.5 yrs versus 60 yrs).
Baseline Median monthly income was USD$90 at
the public hospital and USD$377 in the private
hospital. Of the 94 patients at the 6 month interim
analysis, 18(19%) have died, 19(20%) have been
transplanted, 47 (50%) remain on dialysis and 10
(11%) patients have discontinued dialysis. Median
total monthly expenditure for dialysis was USD$231
in the public hospital and USD$1526 in the private
hospital.
Conclusion
ŠŠ
Establishing a clinical outcomes registry for those
on renal replacement therapy for end stage kidney
disease is feasible and using electronic data capture
tools allows real-time subject follow up, multicenter
participation and centralized monitoring. Our study
found that relatively young Indian dialysis patients
have high mortality and dialysis discontinuation
rates but also a high rate of transplantation. Costs
were high relative to income and are likely to impact
upon ongoing treatment decisions and survival. The
high and ongoing nature of such costs pose particular challenges to how risk protection programs
are designed, particularly given the limited capacity
to pay of its beneficiaries. Our experiences and the
tools developed could be shared within the APAMI
partner institutions many of whom face significant
burden on the health system due to end stage kidney disease.
Keywords: Web-based data collection, Clinical Registry, Open Source Platforms, Clinical Outcomes Monitoring
72
2016 Conference of
Asia Pacific Association for Medical Informatics
STATE OF MANAGEMENT INFORMATION SYSTEM (MIS) IN
TERTIARY CARE HOSPITALS IN PAKISTAN
Suleman Atique1, Abdallah Ahmed Elbakkoush1,
Syed Muhammad Mursalin2, Ting-Wu Chuang4,
Chien-Yeh Hsu3, Shabbir Syed-Abdul1
1_ Graduate institute of Biomedical Informatics, Taipei Medical
University, Taiwan, Taipei, Taiwan
2_ National Institute of Health, Islamabad, Pakistan
3_ Department of Information Management, National Taipei
University of Nursing and Health Sciences, Taipei, Taiwan
4_ Department of Molecular Parasitology and Tropical Diseases,
School of Medicine, College of Medicine, Taipei Medical
University, Taipei, Taiwan
ABSTRACT
Objective
ŠŠ
Objective of this study is to review the existing situation of data collection, processing, analysis and use
of information in major tertiary care hospitals (TCH’s)
across Pakistan. Ultimate aim is to design guidelines
for an effective hospital management information
system (HMIS) compatible with the requirements of
the healthcare system.
Methods
ŠŠ
A cross sectional survey was done in different public
sector TCH’s of Punjab and Islamabad. The method
to collect data involved visits, interviews and distribution of pilot tested questionnaires. Before the
questionnaire distribution it was reminded to each
respondent regarding the questionnaire and asked
about the issues that they might experience. Eight
public sector TCH’s were randomly selected and
contacted. The data obtained was entered and analyzed by SPSS-17.
Results
ŠŠ
Being the provincial capital and the most populous
city of Punjab, 4 hospitals were selected from Lahore. Whereas 2 hospitals were included each from
Rawalpindi and Islamabad. It was observed that all
the hospitals had HMIS/statistical units, responsible
for data collection and analysis. Despite availability
of MIS Unit in all the hospitals there is no MIS management committee found in any. In 6 of hospitals
more than 100 computers were available for the use
of HMIS. Among them, PIMS Islamabad possessed
the maximum number of computers i.e. 300. Whereas in the remaining 2 hospitals this is less than 100.
Specifically for HMIS/Statistical Unit the number
of computers is 15 or more in 6 hospitals, whereas
less than 15 in remaining 2 hospitals. It is found that
out of 8, 4 of the surveyed TCH’s had a proper list
of indicators on which data is being collected. It is
found that supervisors visited the hospitals during
the last three months in four hospitals but no such
visit is recorded made in other four. Regarding what
was discussed during the supervisory visit it was
observed that they discussed hospital performance
and data reports in 4 hospitals but not in others.
Conclusion
ŠŠ
All the TCH’s visited had an MIS Unit adequately staffed. Most of this staff are statisticians and
from IT. Hospital had some kind of data collection
and processing systems, and the efficiency and appropriateness varied among hospitals. There is no
consistency or standardization of data elements
among the hospitals. Each one has developed its
own system without due consideration of others.
Authorities are unable to make any sense out of
this data. In almost all these hospitals there is no
MIS Coordination committee as well as proper data
backup system, with one exception. In general tremendous interest, motivation and was shown both
by Hospital, MIS Units and Health Managers for the
improvement of their MIS. It can be concluded that if
technically supervised, there is sufficient knowhow
and expertise among staff of hospitals for developing a proper MIS. MIS has a lot to offer to improve
healthcare services in Pakistan.
Keywords: Tertiary Care Hospitals. Hospital, Management Information System, Pakistan
73
10
November 4 (Fri), 2016
Oral
Session
Session 10: M
obile Health Application
•
Session 10-1Experience and Preference Factor Analysis for pediatric
obesity management mobile app
- Jisan Lee, Meihua Piao, Ahjung Byun, Hyeonsuk Lee ,
Kyung Ryeon Kwak, Jeongeun Kim
•
Session 10-2Efficacy of Health Management in Rural Area affecting by a
Personal Health Records System with Classification Algorithms
- Hung-Pin Hou, Heng-Shuen Chen, Ching-Chu Liao, Chin-Yao
Chen, Pei-Chia Lo, Mei-Chih Chen, Yu-yuan Yu
•
Session 10-3Analysis of a mobile electronic medical record usage pattern:
User interaction and evolution
- Yura Lee, Jeong Hoon Kim, and Jae Ho Lee
•
Session 10-4Utility of mobile apps in smoking cessation: A literature review
- Shwetambara Kekade, Shabbir Syed Abdul, Yu-Chuan Li
74
2016 Conference of
Asia Pacific Association for Medical Informatics
Experience and Preference Factor Analysis for pediatric obesity
management mobile app
Jisan LEE, RN, M.Sc 1, Meihua PIAO, RN, M.Sc 2,
Ahjung BYUN, RN 2, Lee Hyeonsuk, RN, MSN2,3,
Kwak Kyung Ryeon4, Jeongeun KIM, RN, PhD, INS 1,2,5,
1_ Interdisciplinary Program of Medical Informatics (Seoul National University,
Seoul, Korea)
2_College of Nursing (Seoul National University, Seoul, Korea)
3_ Seoul Women’s College of Nursing (Seoul Women’s College of Nursing, Seoul,
Korea)
4_ Busan Metropolitan Office of Education (Busan, Korea)
5_ Research Institute of Nursing Science (Research Institute of Nursing Science)
ABSTRACT
Objective
ŠŠ
The obesity rate of elementary school students in
Korea (based on body mass index [BMI]) is 20.6% [1].
Smartphone ownership rates are increasing among
children, with 72.2% of elementary school students
in Korea [2]. For an effective pediatric obesity management, some interventions have utilized mobile
technology [3-4]. However, the factors of mobile
design for managing pediatric obesity have not yet
been fully defined. The present study attempted to
clarify the experience and analysis of preference
and needs about mobile app for managing pediatric
obesity (MAMPO).
Methods
ŠŠ
Overweight or Obese high-grade elementary school
students were included as subjects. The researcher
explained the purpose of the study to the principals and the school nurses of schools in Busan. The
nurse-teacher explicated our study’s purpose to students then they took the content to their parents.
Consented participants then completed a questionnaire comprised of children’s experience about
smartphone app and preferred method and UI for
managing weight. We assessed whether a student
wants to participate in making a mobile application
for pediatric obesity management. Statistical analysis was performed using the Mann-Whitney U test
and Kruskal-Wallis H test.
Table1 T
he result of analysis of experience and preference of the overweight and obese elementary school student
MAMPO Partner PrefMethod
experience
erence
Preference
N=350
Student
Grade
Gender
BMI
School
Level
*: p < 0.05
75
Character
Preference
App Preference
Mobile app
Needs
Intention to
use
Z
-0.776
-0.395
-1.632
-2.811
-2.369
-1.642
-1.295
P
0.438
0.693
0.103
0.005*
0.018*
0.101
0.195
Z
-2.793
-2.546
-2.523
-0.971
-4.932
-1.502
-2.732
P
0.005*
0.011*
0.012*
0.331
0.000*
0.133
0.006*
Z
-0.141
-0.346
-0.085
-0.816
-0.168
-0.196
-0.486
P
0.888
0.730
0.932
0.414
0.867
0.845
0.627
χ2
3.696
5.180
1.431
10.758
1.198
8.884
2.108
p
0.296
0.159
0.698
0.013*
0.753
0.031*
0.550
Results
ŠŠ
We recruited 423 overweight and obese students
in 11 elementary schools and 311 students had
their own smartphones. The response rate is 82.7%
(n=350). There are 199 overweight students
(56.9%)/ 151 obese students (43.1%) and 185
Girls (52.9%)/ 165 Boys (47.1%). The participants
consisted of 184 fifth-grade (52.6%)/ sixth-grade
(47.4%) and four school levels (A=9.4%, B=38.4%,
C=46.3%, D=6.0%). Students who want to participate in designing the mobile application for pediatric
obesity are 184(52.5%). Gender mostly influenced
in mobile app experience, partner preference, weight
control method preference, character preference,
mobile app preference, needs of app, and intention to use. Moreover, the results revealed that the
student’s grade and school level showed greater
difference in needs of app, preferred character, and
preferred App (Table 1).
Conclusion
ŠŠ
The purpose of this study is to analyze the elementary school students’ experience and preference for
MAMPO. As a result, gender mostly influenced in experience, preference, needs, and intention to use of
MAMPO. The main strengths of this study are that
we recruited 372 overweight or obese students in
11 elementary schools and revealed their preferred
factors and needs of managing weighty by smartphone. However, there are two limitations about this
study. First, this study was performed only in Busan.
Second, we included the data of whom agreed to
participate in our study. There is a possibility only
positive data were included. In future studies, overweight or obese student’s parents can be included
to compare with their children’s experience and preference. Furthermore, we recommend designing a
mobile service that contains factors which elementary school student prefer.
Keywords: Pediatric Obesity, Mobile Applications, Body Mass Index, Surveys and Questionnaires
76
2016 Conference of
Asia Pacific Association for Medical Informatics
Efficacy of Health Management in Rural Area affecting by
a Personal Health Records System with Classification Algorithms
Hung-Pin Hou 1, Heng-Shuen Chen1,2,
Ching-Chu Liao1, Chin-Yao Chen1,
Pei-Chia Lo1, Mei-Chih Chen 1, Yu-yuan Yu1
1_ Puli Christian Hospital, Puli, Taiwan
2_ National Taiwan University Hospital, Taipei, Taiwan
ABSTRACT
As healthcare services of the aging society become
more complicated, Health management with comprehensive personal information become more and
more important. In Taiwan the high prevalence of
diabetes mellitus and multiple complications is a
major health management issue related to escalating medical cost and complicated problems causing
morbidity and mortality
Objectives
ŠŠ
Because the difficulty of recruiting healthcare
professional at rural area, a personal health records(PHR)system in Puli Christian Hospital was
designed and launched in 2014 to assist case managers in their health management work for achieving
better care quality..
Methods
ŠŠ
In this study, we focused on the analysis of diabetic
cases enrolled in the PHR system using classification algorithms to retrieve patients with abnormal
condition. The system may remind the case mangers
to perform specific intervention according to guideline for standard operation procedure.
Results
ŠŠ
The results showed significant difference of the
HbA1c changes among patients who were enrolled
by the health management PHR system or not
during two years periods from 2014 to 2015. The
patient managed by health management information
system received crucial intervention activities continuously followed by frequent exploration with classification algorithms in their PHR. Besides, another
benefit from the information system was to reduce
paper work and save work loading.
Conclusion
ŠŠ
In this study health management using a PHR system with classification algorithms could effectively
control chronic diseases and provide comprehensive
health management. Especially in rural area intelligent health management system could bridge the
gap of insufficient healthcare professionals.
Keywords: personal health record, health management system, case management, data mining
77
Analysis of a mobile electronic medical record usage pattern:
User interaction and evolution
Yura Lee, MD1, Jeong Hoon Kim, MS2,
and Jae Ho Lee, MD, PhD1,3,
1_ Department of Biomedical Informatics, Asan Medical
Center, Seoul, Korea
2_ Medical Information Office, Asan Medical Center,
Seoul, Korea
3_ Department of Emergency Medicine, University of
Ulsan College of Medicine, Seoul, Korea
ABSTRACT
Objective
ŠŠ
72% of doctors in the United States already own
a smartphone. The high penetration of smartphone
could have thinking about bed side tool in hospital.
A mobile electronic medical record (EMR), called as
mAMIS (Mobile Asan Medical Information System),
was developed at Asan Medical Center, November
2010 and distributed to 900 hospital staffs. Analysis of access logs and service logs was done. We will
analysis the log data of the mAMIS usage pattern,
to figure out user interaction and the direction of
evolution.
Methods
ŠŠ
mAMIS is a mobile EMR viewer which provides electronic medical records via smartphone. The major
services of mAMIS are including patients list (inpatient, emergency department, consult), detailed clinical information (vital signs, blood sugar levels, alert
information), order list, laboratory results, radiologic
results, medication orders and drug information.
Currently, 3,859 clinician and health care providers
are using mAMIS. The logs for the selected services
were collected as described below. The mAMIS logs
were collected from December 2010, when the app
was first released, to December 2015.
half of the users were trainees (59%, 649/1,102).
For daily log from March 2012, total count of users
and log-in were kept increasing, and for the daily usage, the mean user count was 438.3 and mean login count was 1044.9. The service usage pattern of
doctors and nurses are shown in figure 3. In total,
5,233,592 logs and 1,561,216 logs were checked
in nurse group and doctor group. In Both group,
patient lists were frequently used. In nurse group,
services for order information and nurse EMR were
frequently used. In doctor group, services for patient exams and their results were frequently used.
7-8AM and 4-6PM, same times of doctor’s rounds,
were highly accessed times. And in nurse group, two
peaks were shown around 12 pm and 20 pm, one
or two hours before hand-over time of nurses. And
those peaks were corresponded to the peaks of
out-hospital usage.
Conclusion
ŠŠ
A mobile EMR has been used successfully at Asan
Medical Center until now. As the patterns of usage
were correlated with the workflows of different
working group, doctors and nurses, it represents
that mAMIS is working its unique role for efficient
workflow.
Results
ŠŠ
63% of users were nurse (2,333), followed by 30%
of doctors (1,102) and others (pharmacists, healthcare assistants) (Fig.1-a). In doctor group, more than
Keywords: Smartphone application, mobile electronic medical record, usability
78
2016 Conference of
Asia Pacific Association for Medical Informatics
Utility of mobile apps in smoking cessation: A literature review
Shwetambara Kekade,
Shabbir Syed Abdul, Yu-Chuan Li
Graduate Institute of Biomedical Informatics,
Taipei Medical University, Taipei, Taiwan
ABSTRACT
Objective
ŠŠ
Smoking is one of the major causes of preventable
death globally. A number of mobile applications are
currently available to promote health by quitting
smoking. The purpose of the study was to determine
the behavioral changes and the benefits of mobile
applications in smoking cessation.
Methods
ŠŠ
We searched the keyword “mobile apps for smoking cessation” in the following databases: Pubmed,
Sage, Scopus and Science Direct. A total of 348 articles were found. We included those articles which
were published in last five years and described the
mobile applications for smoking cessation. The comparative studies and review articles were excluded.
Results
ŠŠ
After screening of titles, 48 articles were selected,
from which we chose 11 articles following abstract
screening. The duplicate articles were removed and
finally 7 articles were included in the study. Three
articles stated that the mobile app was beneficial in
aiding to smoking cessation, whereas two articles
mentioned about the insignificance of the apps in
quitting smoking; one article stated that the app had
not yet been tested and one article mentioned about
ongoing clinical trial.
Keywords: mobile apps, smoking cessation, quit smoking
79
Conclusion
ŠŠ
Smoking, being a cause of a number of diseases,
numerous apps have been developed in recent years
to encourage its withdrawal. Our study shows that
mobile apps are beneficial for behavior changes and
eventually quitting smoking to some extent. The
apps possess different features such as motivation,
inspiration, avoidance and distraction with an approach to smoking cessation. Although, most of the
apps are found to be user friendly, major issues arise
due to the lack of personal interest and motivation
to quit smoking. A game app has been used for distraction, but it faced the constraint of user frustration. Another reason for the non-beneficial nature of
apps is the small number of individuals or small duration of study. Feasibility of smoking apps depends
upon a number of factors such as its ease to use,
individual urge to quit, the duration of use, motivation and inspiration from others and so on. However,
our study faces the limitations of small number of
articles describing the apps for smoking cessation.
Future studies and clinical trials on a large scale are
suggested for further implications.
11
November 4 (Fri), 2016
Oral
Session
Session 11: Artificial Intelligence in Medical
Informatics
•
Session 11-1Evaluating the effectiveness of distributional hypothesis and
chunk tagging schemes: a case of HMM for clinical named
entity extraction
- Wangjin Yi, Jinwook Choi
•
Session 11-2Coordinated Markov Modeling of Cancer Metastasis from
Multiple Primary Sites
- Hyunggu Jung, Anthony Law, Esther Wu, Mark E. Whipple
•
Session 11-3Structuralization of Variance Text Records in Clinical Pathway
- Takanori Yamashita, Yoshifumi Wakata, Hidehisa Soejima,
Naoki Nakashima, and Sachio Hirokawa
•
Session 11-4Identify Factors Affecting Successful Ageing in Elderly
Diabetics in Shiraz Using Data Mining Approach
- F. Izanloo, M. Nasiri , A. Aslani , M. Hajipour, A. Pakdaman,
H. Ghaem, M. Shafiee
80
2016 Conference of
Asia Pacific Association for Medical Informatics
Evaluating the effectiveness of distributional hypothesis and
chunk tagging schemes: a case of HMM
for clinical named entity extraction
Wangjin Yi1, Jinwook Choi1,2
ABSTRACT
Objective
ŠŠ
This study presents a validation process that helps understand the effect of distributional structure and entity
boundary information in information extraction by implementing a specialized topology in HMM (hidden Markov
model). This study is done for the purpose of verification
of some configurations (i.e., distributional structure,
tagging scheme) considered natural in clinical name
entity extraction domain. The distributional structure is
a form of the distributional hypothesis which suggests
that words are characterized by their neighbor words; in
other words, words sharing similar neighbors tend to be
classified similarly. Representing contextual information
of an entity, the distributional structure is a feature type
which is generally used in machine-learning based name
entity recognition model. Also, BIO tagging scheme
, which marks the beginning of an entity boundary, is
commonly used to represent chunks of an entity in a
text. This paper empirically studies the reason why distributional structure improves the efficiency of information extraction, in addition, evaluates the effectiveness
of different tagging schemes.
Methods
ŠŠ
This study evaluated the performance of multiple topologies of hidden Markov models. Sequence labeling
model is a general choice for a named entity recognition task. Even though conditional random fields, which
is one of sequence labeling models, is empirically proven state-of-the-art method, we chose hidden Markov
model because the parameters of the model are probability matrices and these parameters are interpretable
for a human. So it was easy to be observed how the
distributional structure influences the model. And also,
it is easy to implement.
Observation in both training and test instance was
set as sequences of unigram symbols. The topologies
of HMMs are basically designed by following tagging
schemes such as IO, BIO, and BIEO. IO scheme encodes a boundary of a target entity by a sequence of
a single state (i.e., ‘I’) and uses ‘O’ to represent background information (outside of target entities). BIO
tagging scheme particularly assigns ‘B’ to the beginning
81
1_ Interdisciplinary Program for Bioengineering, Graduate School,
Seoul National University, South Korea
2_ Department of Biomedical Engineering, College of Medicine,
Seoul National University, South Korea
of an entity. BIEO is derived from, which adds ‘E’ tag to
represent the ending of an entity. The topologies are
expanded by adding ‘prefix’ and ‘suffix’ states for the
preceding one and the following one of entity state respectively. Under the setting of using unigram symbol,
we wish that the prefix and suffix state provide context
information which is a distribution of neighbor word and
cannot be implemented in an HMM clique.
In our designed HMM topologies (Fig. 1), transitions
between states are partially connected. The transition
from background to both target entity and suffix state
is suppressed, also, the transition from target entity to
both prefix and background is suppressed. The named
entity recognition experiment was performed on
three named entity recognition task data set; A) i2b2
2010 clinical event identification data set. The data
set consists of 394 electronic discharge summaries
and progress notes which are used for the training set
in i2b2 2010 NER challenge. The notes are obtained
from multiple hospitals in the United States. The target
entities are annotated for medical problems, tests, and
treatments. B) Electronic clinical discharge summaries
obtained from Seoul National University Hospital (SNUH)
in South Korea. The data set consists of 200 discharge
summaries of rheumatic patients who discharged in a
period (2013~2014) from the hospital. The annotated
information is for medical symptoms/signs, tests, diagnosis/disorders, medications, procedures, visit information (hospital, department or verbal (action) expressions).
C: Seminar Announcement information extraction dataset from Carnegie Mellon university (CMU) electronic
board. This data set is used for validation of our model in
a general NER task. It consists of 485 documents and
the annotation includes location, speaker, starting time
(stime), and ending time (etime).
Results
The datasets were split into training and testing parts
(3:1 for i2b2 2010 and SNUH, and 1:1 for CMU).
HMMs were trained by supervised learning. HMMs
were built for each target entity and tagging schemes
(‘IO’, ‘BIO’, ‘BIEO’, and ‘BIOE+prefix/suffix states’), and the
learned models were evaluated on the three datasets.
Table 1 shows the evaluation result (F1-score). According to the result, ‘Prefix’ and ‘Suffix’ states effectively improves the performances in every case. In
particularly, the prefix/suffix states improve Recall
performance. Given the test data that includes an
amount of previously unseen words, the inference
power of the ‘BIEO+prefix/suffix’ model was augmented by the prefix/suffix states.
To verify the reason of the model’s performance improvement by the ‘prefix’ and ‘suffix’ states, we observed the inside of the Viterbi lattice for ‘Speaker’
entity where unseen words are most prevalent. We
synthesized a possible snippet of an observation,
‘lecture who: Arnold’. ‘Arnold’ is the previously unseen
word in test time, and ‘who:’ is a symbol frequently preceding ‘Speaker’ mention. In ‘IO’ tagging scheme, the
symbol, ‘who:’, was estimated as a background symbol,
and it makes the model predicts the symbol, ‘Arnold’,
as a background symbol.
In addition, ‘B’ state representing the beginning of an
entity chunk augmented the emission probability of
symbols which are likely to be emitted from target
state. To verify the reason of the model’s performance
improvement through the tagging schemes, we observed the emission probabilities of words related to
the ‘Location’ and ‘Speaker’ states. The emission probability distribution of randomly selected words in each
model was augmented when the additional states (i.e.,
‘B’, and ‘E’) are added.
Conclusion
ŠŠ
The question posed by this study was how the distributional structure improves the prediction of a named
entity recognition in clinical texts. With specialized
states providing information of the distributional
structure and boundary information, we showed improvements in named entity recognition performance
on two clinical document datasets and one general
information extraction data. The distributional structure scales up the probability of informative words and
makes transitions to correct predictions to be more
favorable in the state transducer. At the same time,
fine-grained target states enlarge the emission probabilities of words that are members of the target entity.
Given that the hypothesis is natural to humans, many
researchers utilize the structure as a baseline feature
in statistical language processing without quantifying
its effectiveness. Hence, we would, based on the findings of this study, like to assert that the effectiveness
of the structure in HMM was proven empirically.
Figure 1 HMM topologies in this study. (a): IO, (b): BIO, (c): BIEO, (d): BIEO+prefix/suffix. Background: non-target state.
Start and Finish states are omitted in this figure.
(a)
(b)
(c)
(d)
Table1 F
1-scores of HMM topologies on the test sets.
i2b2 2010
IO/BIO/BIEO/
BIEO+prefix
SNUH
IO/BIO/BIEO/
BIEO+prefix
Topology
IO/BIO/BIEO/
BIEO+prefix
CMU
IO/BIO/BIEO/
BIEO+prefix
Problem
Test
Treatment
.316/.368/.415/.592
.384/.368/.472/.642
.280/.329/.351/.568
Symptom/sign
Test
Diagnosis
Medication
.342/.404/.479/.575
.597/.631/.632/.722
.502/.555/.562/.643
.750/.752/.752/.831
Procedure
Visit (location)
Visit (action)
.571/.595/.597/.668
.524/.555/.553/.700
.691/.604/.670/.800
Location
Speaker
Stime
Etime
.579/.625/.674/.812
.079/.176/.275/.547
.714/.746/.747/.832
.494/.466/.466/.748
Keywords: Name entity recognition, Clinical event extraction, Text processing, Hidden Markov model
82
2016 Conference of
Asia Pacific Association for Medical Informatics
Coordinated Markov Modeling of Cancer Metastasis from
Multiple Primary Sites
Hyunggu Jung1, Anthony Law2,
Esther Wu1, Mark E. Whipple1,2
1_ Department of Biomedical Informatics and Medical Education,
University of Washington, Seattle, WA, U.S.A.
2_ Department of Otolaryngology-Head and Neck Surgery,
University of Washington, Seattle, WA, U.S.A.
ABSTRACT
Objective
ŠŠ
Planning effective treatment of cancer of the head and
neck requires predicting the probability of microscopic
tumor spread to regional lymph nodes. Previous studies
demonstrated that Markov chain models are feasible
methods to predict tumor spread in the lymphatic system from individual primary tumor sites in the head and
neck. However, little is still known about how to utilize
data from multiple primary sites with overlapping lymphatic drainage to improve model performance. In this
study, we investigated whether a Markov chain model
that was based upon lymphatic drainage pathways
could predict the probability of metastasis to individual
nodal groups. Further, we tested if a Markov model that
uses parameters obtained by training with data from
two different primary sites performed better than a
model trained with data from a single primary site.
Methods
ŠŠ
We created a two-dimensional Markov chain model
where the row of the model represents the metastatic
progression along the lymphatic pathway and the column indicates the T-stages for the primary tumor location. To estimate the parameters of the model, we used
two data sets as training sets: one from the records of
50 patients with non-treated, non-recurrent squamous
cell carcinoma (SCCA) of the oral tongue and 10 patients with SCCA of the buccal mucosa, which presented to the University of Washington (UW) head and neck
tumor board over a 3.5-year period. We ran the model
with all possible parameters with a step size of 0.1 to
identify the upper and lower bound of the parameters
that fit with the training sets. We then determined the
parameters when the output of the model was closest
to the training sets after running the model again for any
parameters with a step size of 0.05 within the range.
For validating the model, we compared the closeness
(i.e., cosine similarity) between the output of the model
and a test set derived from the Cancer Genome Atlas
(TCGA), which included 81 patients with SCCA of the
oral tongue and 9 patients with SCCA of the buccal mucosa. For each of the two primary sites, we compared
the performance between the Markov model trained
with data from the primary site only to the performance
of the model that was trained using data from both
primary sites. For comparison, we also measured the
cosine similarity between the training and test sets.
Results
ŠŠ
The cosine similarity between the output of the
Markov chain model and the test set was greater
than the cosine similarity between the training and
test sets (see Table 1). Also, the output of the model using the parameters trained by combined data
sets from two different primary sites (i.e., buccal
mucosa and oral tongue) showed better prediction
than models trained using a single data set.
Conclusion
ŠŠ
We validated the Markov chain model for predicting
tumor spread using the TCGA data with the estimated
parameters from different primary sites. The results
of our study indicated that our Markov chain model
may accomplish better performance when utilizing
multiple data sets from different primary sites than a
model trained with data from a single primary site.
Table1 Cosine
similarity between two data sets: 1) UW patient data and TCGA data, 2) model output trained with a
single data set and TCGA data, 3) model output trained with combined data and TCGA data
Case for buccal mucosa
Cosine similarity with TCGA
Buccal mucosa data from UW
0.279
Model output trained with buccal
mucosa data
Model output trained with
combined data
0.584
0.587
Keywords: Markov chain, Oral cancer, Metastasis
83
Case for oral tongue
Cosine similarity with TCGA
Oral tongue data from UW
0.889
Model output trained with oral
tongue data
Model output trained with
combined data
0.914
0.967
Structuralization of Variance Text Records in
Clinical Pathway
Takanori Yamashita1, Yoshifumi Wakata1, Hidehisa
Soejima2, Naoki Nakashima1, and Sachio Hirokawa3
1_ Medical Information Center, Kyushu University Hospital, Fukuoka, JAPAN
2_ Saiseikai Kumamoto Hospital, Kumamoto, JAPAN
3_ Research Institute for Information Technology, Kyushu University,
Fukuoka, JAPAN
ABSTRACT
Objective
ŠŠ
Electronic medical data include structured numerical
data and unstructured text data. The medical text
data utilization has not been enough so that it takes
time and effort to need natural language processing
and text mining. The purpose of the present paper is
structuralization of variance text records in clinical
pathway. Clinical pathway has the daily goal and status, medical treatment of inpatient on temporal as the
outcome and unstructured text data such as medical
record and summary is structured efficiently as well
as a structured numerical data such as laboratory
data. However, if cannot be achieved the outcome
or occur some problems, it is recorded by the text of
various expressions about the patient status and the
future plan by doctor or nurse. When those data are
structured or standardized, the variance input operation improve the efficiency.
Methods
ŠŠ
The present paper analyzed about 10,000 of variance text records of fever, which recorded in April
2014 ~ April 2016 at Saiseikai Kumamoto Hospital.
First, it performed morphological analysis to those
data with the medical dictionary of 70,000 word,
then it was applied N-gram(N=2~5) to those words.
All the words of each morpheme word and morpheme
2~5grams as attributes were applied hierarchical
clustering by Cluto, then it drew sparse matrix with
row dendrogram of the word and column dendrogram
of document. The combination of the attributes of
the row and column is a strong portion is visualized
emphasized in sparse matrix. The number of clusters
in rows specified 10.
Result
ŠŠ
There were many attributes overlapped semantically,
therefore it gathered similar attributes in the sparse
matrix. It was extracted feature sentences ("follow-up", "cooling", "analgesic", "check infection") based
distributions of the word group in sparse matrix. The
sentence about "follow-up" was most often appeared.
There were described the plan of starting and continuation or the site about "cooling", administration and
its effect about analgesic, and related to pneumonia
or urinary tract infection about "check infection".
Conclusion
ŠŠ
We have analyzed to medical text records and summary by text mining for the purpose of medical care
process improvement, however clinical interpretation
was difficult by the only morpheme word. Although
some of the Japanese-specific issues that is difficult
to separate word, it appeared similar words or sentences as a group in the sparse matrix by using all the
words of morpheme N-gram. We consider that has
been progressing method for interpretation. This time
the target was a variance record of fever. Then, we
plan to classify extracted words for each case, and
expand the target to other variance records (e.g. pain
control, dietary intake, blood pressure).
Keywords: clinical pathway, text mining, N-gram.
84
2016 Conference of
Asia Pacific Association for Medical Informatics
Identify Factors Affecting Successful Ageing in Elderly Diabetics
in Shiraz Using Data Mining Approach
F. Izanloo1, M. Nasiri 1, A. Aslani 1, M. Hajipour2,
A. Pakdaman1, H. Ghaem3, M. Shafiee4
1_ Health Human Resources Research Center, School of Management &
Infor mation Sciences, Shiraz University of Medical Sciences, Shiraz, Iran
2_ Department of Epidemiology, School of Health, Shiraz University of
Medical Sciences, Shiraz, Iran.
3_ Assistant Professor Research Center for Health Sciences, Epidemiology
Department, School of Health, Shiraz University of Medical Sciences, Shiraz, Iran
4_ Department of Nursing, School of Nursing and Midwifery, Shiraz University
of Medical Sciences, Shiraz, Iran.
ABSTRACT
Objective
ŠŠ
In recent decades, world population rapidly became
old. In low and middle-income countries population
change is more rapid. In Iran, older people above 60
years contains 8.2% of the whole country’s population, and furthermore, It’s predicted that it will reach
10% by 2020 which shows the speed of aging in
this country. Diabetes can have an effect on successful aging, but there is not enough information
about the relation between diabetes and successful
aging. In this study, we use data mining approaches
to extract important factors about successful aging
in diabetics in Shiraz.
Methods
ŠŠ
In this study, information of 592 older people (aged
≥60 years) with diabetes between April 2014 and
April 2015 are collected from health centers of
Shiraz and then we used the APRIORI algorithm
for analyzing them. The method of collecting information was a questionnaire contains 99 features
from these domains: Physical health from Bartle
questionnaire, demographic information, important
diseases, quality of life, life satisfaction from Diner
questionnaire, social health from Keyes questionnaire, spiritual health from Alison questionnaire. We
used association rule to analyze the relationship
between features and identify important features
influence successful aging.
Keywords: aging, successful aging, diabetes, data mining
85
Results
ŠŠ
Results of analysis of the relationship between every
single domain and successful aging show that every
domain has a relation with successful aging. The
association rule algorithm applied on all factors that
show 32 rules achieved 7% of support and a confidence of 50%-51%. In the last model, the majority
of rules was about religious health.
According to the results, the most important factor
influenced successful aging was spiritual health
while religious health was more important than existential health. After these, social health and factors
like a feeling of valuable for society, being able to
make some valuable product for society and quality
of life (pain, sleep disorder and the food must not eat
due to diabetes) in older people with diabetes were
more important than other factors. Finally, after all
of these, diabetes, liver disease, kidney disease and
CVA had more impact on successful aging than other chronic diseases.
Conclusion
ŠŠ
In this analysis, religious health was the most important factor, after that social health and quality of life
were important. These factors had subdomains that
priority of them determined. Results of this study
consistent with findings of other researchers about
effective factors on successful aging. Current study
with analyzing of all factors provided the possibility
of identifying more important factors.
P1
November 3 (Thu), 2016
Poster
Session
•
Online Medical Appointment System
•
A comparison of community-based tele-rehabilitation, face-to-face, and usual care for
low-income elderly women in Korea
•
Hyoid Bone Tracking using Tracking-Learning-Detection in Videofluoroscopy Swallowing Study
•
Characterizing hidden rules linking symptoms and acupoint selection using an artificial neural
network model
•
Deep Belief Network based Classification Prediction Model of Pathology Stage in Patients with
Prostate Cancer
•
Developing novel algorithms of approximate entity extraction for laboratory test names in
clinical documents based on q-gram
•
Development of new extravasation alert system by optical hemoglobin sensor in drip infusion
therapy
•
Improving of Medical Supply Purchasing Process by Using Barcode Technology in Thai
University Hospital
•
AURORA: Analytic code managing and distributing Unit bridging Research Organizations and
Associated users
•
Exploring the possibility of identifying adverse-drug-events from nursing records
•
A method of compressing whole slide images (WSI) with multi-focus plane using blur estimation
•
Suggestion of a Phased Classification Method of Requirements for a Registry by the types of
Cancer and the Definitions of Items using Electronic Medical Records
•
Development of Patient Portal for Exchange Personal Health Record
•
FHIR for IHE LAW Profile
•
Design of middleware solution for the vendor neutral archive system
•
Design and Implementation of SMART guide system based on integrated data modeling
•
Design and Development of a Medical Information Support System
Keywords: aaaaa
86
2016 Conference of
Asia Pacific Association for Medical Informatics
Online Medical Appointment System
Winnie Hoi Wing Leung
Master of Information Systems (University of Melbourne)
ABSTRACT
Objective
ŠŠ
Business process reengineering represents an important opportunity for business to reduce cost and
increase quality at proper time and speed. The purpose of process redesign is to identify possibilities
for improving the design of medical appointment systems: from "as is" to "to be." The aims of this process
redesign are to increase healthcare quality, efficiency,
flexibility and decrease cost.
Method
ŠŠ
The proposed medical appointment system is designed for small or medium size medical centres (clinics) with a reception assisting the user to administer
data sets. Make / modify appointment is the chosen
target process for this project. I improve the design
process by using the Business Process Modelling and
Notation (BPMN). I analyse the strengths and weaknesses using several heuristic design principles and
the devil’s quadrangle model from Dumas (2013).
Result and Conclusion
ŠŠ
I’ve improved the design process by introducing an
online medical appointment system. It has a highly
efficient set of management tools to synchronise,
computerize and systematically record data assisted
by the use of Internet websites. The medical appointment system based on BPMN describes the process
of patients making medical appointments using both
traditional method and online system.
There are at least eight principles identified in this
target process redesign: Control relocation, Outsourcing, Integration, Parallelism, Activity automation, Integral technology, Interfacing, Centralization,
Empowerment. Three principles out of the eight are
chosen as examples from the first three groups of
principles, as summarised in Table 1:
• Control relocation – Patients take control over
which doctor and the time for the appointment.
It improves flexibility for the patient as they can
make an appointment online without restricting
to business hours. It also improved patient satisfaction. However, there's a high probability of
patient misusing the system.
• Parallelism - Activities can be carried out in parallel especially when greater speed is required
during the busy business hour. Overall, parallelism leads to improved performance, reduction of
waiting times, reduce throughput time and better
use of capacity.
• Activity automation/ Integral technology -Online
medical appointment technology is deployed to
alleviate physical constraints in a business process. It reduces the time that the receptionist
spends on electronic work. The online system provides a better quality of service. Activities can be
executed faster, increase communication speed,
increase information availability, reduce duplicated data entry, reduce human error and offer a
more predictable result. However, the cost of implementation, training, and maintenance efforts
related to technology can be high. Workers' reluctancy to adopt new technology may decrease the
quality of the business process. Another risk is
that the Internet can be unresponsive, resulting in
a failure to make an appointment.
I’ve analysed the three redesign principles on each
of the four performance dimensions of Devil’s Quadrangle. The dimension of time, quality and flexibility
will improve dramatically. The biggest drawback is
increasing in cost. However, utilising other redesign
principles identified in this medical appointment system, the possibility of improving the Devil’s Quadrangle may occur.
Table1 Characteristics
of the business process operation heuristics
87
Time
Cost
Quality
Flexibility
Control relocation
Neutral
Negative
Positive
Positive
Parallelism
Positive
Negative
Neutral
Negative
Activity automation
Positive
Negative
Positive
Negative
Integral technology
Positive
Negative
Positive
Positive
A comparison of community-based tele-rehabilitation,
face-to-face, and usual care for low-income elderly women in Korea
Hanna Choi1, Jeongeun Kim2
1_ Seoul National University Hospital, Seoul, Korea
1, 2_ College of Nursing, Seoul National University, Research
institute of nursing science, Seoul, Korea
2_ College of Nursing, Seoul National University, Seoul, Korea
ABSTRACT
Objective
ŠŠ
Nearly all low-income older women have a chronic
disease; they are more likely to face them than average elderly people. However, these concerns rarely
receive proper care or further advanced rehabilitation. Recently, telerehabilitation has been recommended as a possible solution to access barriers
with low cost, functionality, and social surroundings.
The objective of this study was to explore how community-based technology with theraband can improve cognitive and physical function, compared to
face-to-face or usual care.
Methods
ŠŠ
Based on a quasi-experimental nonequivalent pretest and post-test research design, 61 low-income
elderly community-dwelling women participated in
three groups: the experimental group, which included
community-based tele rehabilitation with theraband
(n=22); control group 1 with face-to-face theraband
rehabilitation (n=18), and control group 2 with usual
care and free gymnastics (n=21). Participants were
recruited from communities in Seoul and Kwangju,
Republic of Korea. The intervention occurred twice
per week, with each lasting 40 minutes for a period
of 12 weeks. The degree of cognitive variables such
as falls efficacy, exercise efficacy, and vitality was
measured using structured questionnaires. Performance-based measures of physical function data,
SPPB (Short Physical Performance Battery: balance
score, gait speed score, and standing from a chair)
were collected from December 2014 to March
2015. Statistical analyses included a one-way ANOVA and a χ2 test.
Results
ŠŠ
The results were as follows: falls efficacy, exercise
efficacy, and vitality between groups were significantly different (F=8.03, p<.001). Cognitive variables of the experimental group were found to be
significantly higher than that of control group 2
(p<.001). However, there was no difference between
control group 1 and the experimental group. The
three variables of control group 1 were significantly
higher than those of control group 2 (p<.001). Secondly, there was no difference between the total
SPPB for monitoring physical functioning score of
the experimental group and control group 1; however, the SPPB total score of the experimental group
was significantly higher than that of the control
group 2 (F=5.45, p=.007). The SPPB total score of
control group 1 was significantly higher than that of
control group 2 (p<.001). Thirdly, SPPB gait speed
measured of the experimental group was higher than
that of the control groups (p<.001); no difference
between control group 1 and 2 was observed.
Conclusion
ŠŠ
This study shows that telerehabilitation with theraband improves cognitive and physical performance.
Community-based telerehabilitation with theraband
was effective and feasible in a class with face-toface intervention in low-income women. Therefore,
Findings indicate that telerehabilitation be used
to encourage for better assist elderly women with
low-income.
Keywords: Telerehabilitaion, low income, elderly, women
88
2016 Conference of
Asia Pacific Association for Medical Informatics
Hyoid Bone Tracking using Tracking-Learning-Detection
in Videofluoroscopy Swallowing Study
Dongheon Lee1, Jung Chan Lee2,3,4, Byung Mo Oh5,
Hyoun-Joong Kong6,7 and Hee Chan Kim2,3,4
1_ Interdisciplinary Program for Bioengineering, Graduate School, Seoul
National University, Seoul, Korea
2_ Department of Biomedical Engineering, Seoul National University College
of Medicine, Seoul, Korea
3_ Department of Biomedical Engineering, Seoul National University Hospital,
Seoul, Republic of Korea
4 _Institute of Medical and Biological Engineering, Medical Research Center,
Seoul National University, Seoul, Korea
5_ Department of Rehabilitation Medicine, Seoul National University College
of Medicine, Seoul, Korea
6_ Department of Biomedical Engineering, Chungnam National University
College of Medicine, Daejeon, Korea
7_ Department of Biomedical Engineering, Chungnam National University
Hospital, Daejeon, Korea
ABSTRACT
We present a new approach for tracking the hyoid
bone in Videofluoroscopy Swallowing Study (VFSS).
In the current approach, it is difficult to automatically localize the cervical vertebrae from VFSS images
because of occlusion due to a marker used to display the relative position of the cervical vertebrae
region. Moreover, traditional image processing approaches such as template matching or optical flow
methods are not robust for tracking the hyoid bone
in video frame images. In this study, an Active Shape
Model (ASM)-based fitting algorithm is proposed to
detect C2, C4 points and address the aforementioned issue of occlusion. Then, the Tracking-Learning-Detection (TLD) algorithm is applied to automatically track the hyoid bone region in the images. Our
approach is evaluated for the VFSS dataset of four
Figure 1 Cervical vertebrae localized using an ASM
89
dysphagia patient cases; the obtained results show
a total mean error of 13.3±9.05 pixels in comparing
manual trajectory of hyoid bone. Therefore, our proposed method is robust for detecting and tracking
the hyoid bone automatically.
Objective
ŠŠ
Development of a tracking algorithm for the movement of the hyoid bone in VFSS.
Methods
ŠŠ
The proposed method consists of two steps. First,
an ASM[1] for the detection of cervical vertebrae,
specifically C2 and C4, which are used for calibrating hyoid bone coordination, is applied to the
preprocessed initial frame (Figure 1). Second, the
Figure 2 The hyoid bone detected using the TLD algorithm
TLD algorithm[2] was applied to track the hyoid
bone (Figure 2). The method is validated using the
obtained VFSS dataset from four dysphagia patient
cases by comparing the coordination errors of the
hyoid bone, which were calculated using the Root
Mean Square Error (RMSE) between the coordinate
of the hyoid bone in the proposed method and the
reference coordinate.
Conclusion
ŠŠ
This study demonstrated the use of ASM for detecting C2, C4 regardless occlusion between the cervical vertebrae and the marker and TLD algorithm for
tracking hyoid bone. Through our obtained results,
our proposed method is feasible for detecting and
tracking the hyoid bone automatically.
Results
ŠŠ
The results obtained from the evaluation of
our proposed method had a total mean error of
13.3±9.05 pixels in calculating the hyoid bone
trajectory from the VFSS images; the size of
these images was 480×480 pixels and the frame
rate was 30fps. (Table1)
Table1 Errors
in tracking the hyoid bone for the four test subject cases
Mean±S.D (pixel)
Total Mean±S.D (pixel)
Patient #1
Patient #2
Patient #3
Patient #4
8.92±4.78
12.83±7.42
15.69±11.57
15.77±12.43
13.3±9.05
Keywords: Dysphagia, Hyoid Bone, Video Processing, Active Shape Model, Tracking-Learning-Detection
90
2016 Conference of
Asia Pacific Association for Medical Informatics
Characterizing hidden rules linking symptoms and acupoint
selection using an artificial neural network model
Won-Mo Jung1, In-Soo Park1,
Bo-Hyoung Jang2, Younbyoung Chae1
1_ Acupuncture & Meridian Science Research Center, College of
Korean Medicine, Kyung Hee University, Seoul, Republic of Korea
2_ Department of Preventive Medicine, College of Korean Medicine,
Kyung Hee University, Seoul, Republic of Korea
ABSTRACT
The Abstract must be submitted in Microsoft Office
Words and is limited to 3000 characters. If it is necessary, supporting figures can be presented as part
of the text and total content is limited to one page
including supporting figures.
Methods
ŠŠ
In total, 232 clinical records of acupuncture treatment were collected using the “Charting Language”
program. Using an artificial neural network (ANN), we
trained associations between symptom information
and selected acupoints for the treatment. We chose
11 hidden nodes with the highest average precision
score through 10-fold cross-validation.
Results
ŠŠ
Our ANN model could predict the selected acupoints based on symptom and disease information
with an average precision score of 0.865 (precision:
0.911; recall: 0.811). We visualized the relationships
between symptom information and selected acupoints based on learned treatment pattern of hidden
nodes. Two distinctive patterns were observed: 1)
hidden node mainly involved in the regional control
effects of acupoints involved in the treatment of
musculoskeletal pain. 2) hidden node mainly involved
in the remote control effects of acupoints for more
systemic symptoms.
Conclusion
ŠŠ
Objectives
ŠŠ
To understand the principle underlying selection of
appropriate acupoints, it is important to understand
how medical doctors diagnose and treat diseases.
We explored the pattern recognition process pertaining to symptoms and diseases that in informs
acupuncture treatment in a clinical setting.
Our ANN may be a useful tool for diagnostic classification or pattern recognition and for prediction and
modeling of acupuncture treatment based on clinical data obtained in a real-world setting. The relationship between symptoms and selected acupoints
could be systematically characterized through
knowledge discovery processes
Keywords: Acupuncture, indication, neural network, pattern identification, prediction
91
Deep Belief Network based Classification Prediction Model of
Pathology Stage in Patients with Prostate Cancer.
Jaekwon Kim1, Mi Jung Rho2, Jong Sik Lee1,
Young Hyun Pack3, Ji Youl Lee3 and In Young Choi2
1_ Department of Computer Science and Information Engineering, Inha
University, InhaRo 100, Nam-gu, Incheon, South Korea
2_ Department of Urology, Seoul St. Mary’s Hospital, College of Medicine,
The Catholic University of Korea, Seoul, Republic of Korea
3_ Department of Urology, Seoul St. Mary’s Hospital, College of Medicine,
The Catholic University of Korea, Seoul, Republic of Korea
ABSTRACT
Objective
ŠŠ
Prediction model for prostate cancer of pathology
staging is estimating the likelihood that the cancer
has spread before treatment. Although important
for determining the most suitable treatment for patients, pathology staging continues to present significant challenges to clinicians. Clinical test results
such as the pre-treatment PSA (Prostate-Specific
Antigen), Gleason score in tissue biopsies, and the
clinical T stage can be used by clinicians to predict
the pathology stage of prostate cancer. However,
not most patient will return abnormal results in most
of tests. This significantly influenced the capacity to
effectively treatment of prostate cancer. Therefore,
deep learning based pathology stage of prediction is
required to the treatment of prostate cancer.
Methods
ŠŠ
We proposed deep belief network based pathology
stage classification prediction model. Proposed
classification prediction model aim at classify the
OCD (Organ-Confined Disease; pT2+) and NOCD
(Non Organ-Confined Disease; pT3+ or N+). The
study data comprised the data of 467 male patients
extracted from SPCDB (Smart Prostate Cancer
Data Base). Among 467 patients with OCD is 235
patients, NOCD is 232 patients. The DBN model input following variables: PSA, Gleason score, clinical
t stage. DBN consist of two RBM (Restricted Boltzmann machine) layer, and learning to use use a back
propagation algorithm.
Results
ŠŠ
The experiment divided into learning dataset (70%,
OCD: 328 patients; NOCD: 161 patients) and validation dataset (30%, OCD: 141 patients; NOCD:
71 patients) to measure performance. Experiential
result, proposed DBN model accuracy (82.00%)
was larger than the Partin table (66.43%), artificial
neural network (81.43%) and logistic regression
(80.71%).
Conclusion
ŠŠ
The proposed DBN model was superior to the Partin tables in terms of predicting the risk of prostate
cancer. Compared to the validation of the Partin
tables, the DBN model resulted in a larger accurate
prediction of the pathologic stage of prostate cancer.
Keywords: Prostate cancer, Pathology stage prediction, SPCDB, Deep Belief Network
92
2016 Conference of
Asia Pacific Association for Medical Informatics
Developing novel algorithms of approximate entity extraction for
laboratory test names in clinical documents based on q-gram
Kyungmo Kim1, Jinwook Choi M.D., Ph.D.1,2
1_ Interdesciplinary program for Bioengineering, Seoul National
University, Seoul, Korea
2_ Dept. of Biomedical Engineering, College of Medicine, Seoul
National University, Seoul, Korea
ABSTRACT
Objective
ŠŠ
In clinical documents, Entities are often expressed
in a short term. For example, "Neutrophils" is written
as "neuts". We developed a novel approximate entity extraction algorithm that can extract this kind
of expression in clinical documents. We suggested
space effective, less time-consuming algorithm at
the same time.
Methods
ŠŠ
We expressed q-grams on two dimensions. (x-axis :
the number of q-grams of a word, y-axis : the number of matching q-grams). In counter filtering, words
that have LB at least are treated as similar words.
(LB = max(|word|,|query|)-q+1-q*τ, τ is edit distance threshold). So we can think that word that has
matching grams more than y=LB can be candidate
set. We set this methodology as the baseline. And
then we developed model following steps: Method1:
words on the line are components of a set that Edit
distance is 1. For example, if a query word is "neutrophils", there can be "neutrophil", "neut" etc. These
words are treated as exemplary words as short term
in clinical notes. So the points that are closer to
the line are similar to the query. Method2: Different
thresholds are applied to each point upon to their
L. In figure 1, Each triangle has the same area. So
words that are far from the point of query word are
applied strict d(threshold) than nearby words. A Big
triangle is integral space of small triangles so the
points within this area are treated as similar to the
query word. Method3: We added a parameter to
make the big triangle area distorted to exclude some
useless expressions. This model doesn't iterate over
Inverted index of q-gram, instead, just remember
numbers that the number of q-grams(x-axis) and
matching q-grams(y-axis), then calculate Euclidean
distance. So it is space effective, is able to achieve
O(1). We extracted candidate words from 300 doc-
uments of i2b2 2014 that consists of 25,823 tokens with repetition removal. Query words consists
of "Alanine", "cholesterol", "Microalbumin", "troponin",
"Globulin", "Neutrophil", "Lymphocyte", "Monocyte",
"Eosinophil", "Basophil", "glucose".
Results
ŠŠ
Nevertheless extracting additional short terms in
addition to query terms, proposed model 2, 3 could
generate smaller size of candidate set compared to
baseline within every τ dramatically .
Conclusion
ŠŠ
In this study, we proposed a novel model to extract
similar words in clinical documents utilizing Euclidean distance.
Figure 1 Model 3
Figure 2 candidate set
Keywords: Information retrieval, computing methodologies, similar entity extraction, q-grams, abbreviation
93
Development of new extravasation alert system by optical
hemoglobin sensor in drip infusion therapy
Hiroyuki Nozaka
Hirosaki University Graduate school of health
sciences, Hirosaki, Japan
ABSTRACT
Introduction
ŠŠ
Drip infusion is the most popular medical treatment.
Automatic infusion pump system has the flow rate
control or jagging detection mechanism for medical incident prevention, however, it is reported that
medical incidents are often caused in drip infusion
therapy. The cause of the incident is the extravasation from blood vessel, and it is impossible to detect
by the current system. The aim of this work is development of the extravasation detection system
and monitoring system for nursing support in drip
infusion therapy. We examined the application as the
new extravasation alert system.
System configuration
ŠŠ
This system is composed of two photo-reflector
units, a spectroscopic sensor and data logger system "NI-Compact DAQ". Two photo reflector units
are put on the vessel, and it detects blood hemoglobin optically. On the other hand, a spectroscopic
sensor is put on the skin around the infusion position, it scan the maximum wavelength from 400nm
to 900nm. Application software was developed by
"NI-LabVIEW".
Results
ŠŠ
The optical hemoglobin method and the conventional method showed correlation, but optical hemoglobin method was not suitable for absolute quantification of hemoglobin concentration. The intensity of
reflected light was changed significantly between
the upstream and downstream site. The rate of
change of the reflected light intensity showed a correlation to blood dilution by drip infusion.
Conclusion
ŠŠ
The optical hemoglobin method is not suitable for
absolute quantification of hemoglobin concentration. However, this method is suitable for relative
quantitation of hemoglobin, it is considered to be
applicable to new extravasation alert system for
medical safety in drip infusion therapy.
Figure 1 System conception diagram
System evaluation
ŠŠ
Each photo-reflector was placed on blood vessel
at the upstream and downstream site of the needle
injection position (Fig.1). We measured the initial
value of the reflected light intensity on both sides
before infusion. After drip infusion starting, the rate
of change of the reflected light intensity was measured. Additionally, peripheral blood was drawn from
the downstream site continuously. We analyzed the
signal, and evaluated correlation between transcutaneous optical hemoglobin method and the SLS-hemoglobin method.
Keywords: Drip infusion, extravasation, medical incident prevention, nursing support
94
2016 Conference of
Asia Pacific Association for Medical Informatics
Improving of Medical Supply Purchasing Process by
Using Barcode Technology in Thai University Hospital
Riponchai L, Pongskul C, Jiravuttipong A,
Bhoomibunchoo C
Faculty of Medicine, Khon Kaen University, Khon Kaen, THAILAND
ABSTRACT
Srinagarind Hospital, a 1000 beds university hospital implemented new procurement software solutions in June 2015, the hospital had already implemented Hospital Information System (HIS) since
December 2014. Data was cleaned and new item
codes was set and used (>17,000 medical supply items). The new purchasing process took more
time to finish due to longer item code names. We
conducted a pilot project of barcode systems implementation in our operating theater to evaluate the
benefit and value of this technology in purchasing
process.
Objectives
ŠŠ
To compare using of barcode technology in the purchasing process of medical supplies
Methods
ŠŠ
We conducted a comparison experiment between
manual data input and barcode assisted system
during purchasing process of 40 medical supply
items in operating theater unit. We assessed the
errors, processes reduction and time reduction in
purchasing order processes.
Keywords: Purchasing, barcode, operating theater
95
Results
ŠŠ
data input error was found 3 times by manual data
input and 0 by barcode methods. Process step was
reduced from 7 to 3 steps after barcode system
was implemented. Mean total process time was
36.94 seconds/step by manual data input and 6.49
seconds/step by barcode systems.
Conclusion
ŠŠ
barcode application can reduce data input error,
medical supplies purchasing process and total process time.
AURORA: Analytic code managing and distributing Unit bridging
Research Organizations and Associated users
JungHyun Byun, B.E., Hojun Park, B.E., MinSeok Jeon, B.E.,
Sungjae Jung, B.E., Dukyong Yoon, MD,
Ph.D, Rae Woong Park, M.D, Ph.D
Department of Biomedical Informatics, Ajou University
School of Medicine, Suwon, Korea
ABSTRACT
We need to distribute an analytic code to the data
partners to perform a study using the Distributed
Research Network (DRN). Although we have successfully performed many analyses by using email to
communicate and distribute the analytic codes, however we need a more sophisticated managing tool
for them. We develop a managing and distributing
tool for the analysis codes between the researcher
and data partners and named Analytic code managing and distributing Unit bridging Research Organizations and Associated users (AURORA). With the
AURORA, a researcher can easily send their study
protocol and analytic codes to multiple data partners and check the status of the study progress.
Objectives
ŠŠ
Distributed research network (DRN) has been
emerging as an alternative mean for maintaining
confidentiality among analysis using information
from multiple institutions. To perform an analysis
with many data partner, we need to send the analysis codes to many data partners. Although we have
successfully performed many analyses by using
email to communicate and distribute the analytic
codes, however we need a more sophisticated managing tool for them. We tried to develop a managing
and distributing tool for the analysis codes between
the researcher and data partners.
Methods
ŠŠ
Design of the system
We named the system as Analytic code managing
and distributing Unit bridging Research Organizations and Associated users (AURORA). Through the
system, a researcher can send a research protocol
and analytic codes to multiple data partners via
SMTP. Each administrator of the data partner can
determine to approve study or not. For the approved
study, the analysis is performed in the organization.
The decision, comments from the administrator, or
the results of the analysis are automatically provided to the researcher. Delivery of messages between
researcher and administrator of each data partners
is performed by e-mail (e.g., sending code, notifying
the change of status)
Distribution and production of the web interface :
Using Java Database Connectivity (JDBC) API, we
let the system be able to access the database and
record the log. The system send all of the events
that occur during the process of the study by e-mail
to the researchers and the administrators of the
organizations, to enable real-time monitoring of the
research.
Results
ŠŠ
Only authorized user can access the system. The
system has the following functions: uploading and
downloading for protocol, analytic code and result
files; mailing for alarm and link; status update for job
and error; member administration.
Conclusion: With the AURORA, a researcher can
easily send their study protocol and analytic codes
to multiple data partners even without information
who is in charge of the system. Each institution can
confirm the study protocol and perform the analysis by the received analytic code. Eventually, each
administrator of data partner can send the results
from the analysis to the original researcher through
the AURORA. Since only the result of the analysis is
requested and transmitted, the personal information
is thoroughly protected, and the anonymity is maintained.
Keywords: : Distributed Research Network, Manage System
96
2016 Conference of
Asia Pacific Association for Medical Informatics
Exploring the possibility of identifying adverse-drug-events from
nursing records
Eunjoo Jeon 1, Hyeoun-Ae Park 1,
Eunja Chung 2, Hyunok Bae 3
1_ School of Nursing (Seoul National University, Seoul, Republic
of Korea),
2_ Department of Nursing &
3_ Team for Medical Informatics (Seoul National University
Bundang Hospital, Seongnam, Republic of Korea)
ABSTRACT
Objective
ŠŠ
OBJECTIVE: The purpose of this study is to explore
whether the drug contraindications and adversedrug-events data can be extracted from the nursing
records.
Methods
ŠŠ
Two nurses analyzed the drug information available
from KFDA [1] of ten randomly selected drugs [2]
for the contraindications and adverse events. Then,
eight nurses were asked to check where these 334
extracted contraindications and adverse events can
be identified either in nursing records, medical records or lab reports.
Results
ŠŠ
As shown in Table 1, nursing records have more information on the contradictions and adverse events
in ‘Psychiatric’, ‘Cardiovascular’, ‘Respiratory’, ‘Body
as a whole’, ‘Application site’, ‘Neurological’, ‘Skin and
appendages’ and ‘Gastrointestinal’ categories compared to the medical records or lab reports. Medical
records have more information on the factors in
‘Vision’, ‘Endocrine’, ‘Infections’, and ‘Personal’ categories than the nursing records and lab reports do.
Lab reports have more information on the factors
in ‘Musculoskeletal’, ‘Metabolic and nutritional’, ‘Liver
and biliary’ and ‘Blood’ categories than the nursing
and medical records do.
Conclusion
ŠŠ
This study showed that the possibility of using
nursing records as a source for detecting the early
signals for the adverse drug events, especially psychiatric, neurological, cardiovascular, and general
symptoms.
This work was supported by the National Research
Foundation of Korea (NRF) grant funded by the Korea government (NRF-2015R1A2A2A01008207).
Table1 Types of records where contraindications and adverse-events of ten drugs identified by categories
N of
Sx
Nursing
records
Medical
records
Lab
report
Category
N of
Sx
Nursing
records
Medical
records
Lab
report
2
2(100)
2(100)
0( 0 )
Gastrointestinal
26
20(76.9)
9(34.6)
4(15.4)
3
3(100)
3(100)
0( 0 )
Vision
8
5(62.5)
7(87.5)
0( 0)
7
7(100 )
2(28.6)
0( 0 )
Urinary tract
12
6(50.0)
3(25.0)
6(50.0)
16
16(100 )
0( 0 )
1(6.25)
Endocrine
9
4(44.4)
6(66.7)
5(55.6)
Respiratory
13
Body as a
8
whole
Application
3
site
Neurological
37
Skin and append27
ages
Drug
6
13(100 )
1( 7.7)
3(23.1)
10
4(40.0)
4(40.0)
6(60.0)
8(100 )
0( 0 )
0( 0 )
Musculoskeletal
Metabolic and
nutritional
22
5(22.7)
3(13.6)
19(86.4)
3(100 )
0( 0 )
0( 0 )
Infections
20
4(20.0)
20(100)
17(85.0)
36(97.3)
5(13.5)
1( 2.7)
Personal
71
14(19.7)
67(94.4)
9(12.7)
24(88.9)
5(18.5)
0( 0 )
Liver and biliary
10
1(10.0)
1(10.0)
10(100)
5(83.3)
6(100)
0( 0 )
Blood
24
2( 8.3)
4(16.7)
24(100)
Category
Immune
Hearing and
vestibular
Psychiatric
Cardiovascular
Keywords: adverse drug event, drug surveillance, electronic medical record, electronic nursing record.
97
A method of compressing whole slide images (WSI) with
multi-focus plane using blur estimation
Manwon Hwang, Dong Ouk Kim, Sungyong Kim,
Pham Duy Hien, YongSeok Kim
R&D Center, INFINITT Healthcare Co.,Ltd. Seoul
ABSTRACT
When scanning a lab specimen for digital pathology,
some region of the specimen may come out blurry
when its curved surface gets out of focus. As this hinders clear visualization of the area of interest, some
scanners are designed to perform multiple scans,
adjusting focal lengths, which results in creation of
many images. Among the multiple images, some
would eventually show the region of interest vividly.
However, such a method is time-consuming and ends
up with large data size. As high-volume data leads to
high cost of storage and transmission, we propose a
new method to improve the issue.
This paper explains how to compress whole slide
images (WSI) by creating virtual images to reduce
image size without compromising on image quality:
compare different slide images of a specimen created by adjusting focus to identify the sharpest image
for each region of the image on a given 2D block,
then estimate similar blurs with the rest of the image to create a virtual image. Processing and saving
only the difference between the estimated virtual
image and the original image results in substantial
decrease in data size, without affecting its quality.
The presented method is expected to allow effective data size reduction for long-term storage, as
well as significant saving in data transmission cost.
Keywords: Digital Pathology, WSI, Compression, Blur Estimation, Multi Focus Plane
98
2016 Conference of
Asia Pacific Association for Medical Informatics
Suggestion of a Phased Classification Method of Requirements
for a Registry by the types of Cancer and the Definitions of Items
using Electronic Medical Records
Yeong Mi Jang, Il Won Sin ,Kyung Ha Lee ,
Hyo Soung Cha and Seung Hyun Chung
National Cancer Data Center, National Cancer
Center, Go-Yang, Korea
ABSTRACT
Objective
ŠŠ
Although medical records are well computerzied in
these days, the utilization of medical information
data is low. This is one of reasons that general hospitals build up research registry for meaningful use
of the electronic medical records (EMR). Due to the
fact that existing requirements of registry by cancer
type and methods of the definitions of item varies by
hospitals, it is necessary to organize classification
standard.
Method
ŠŠ
This study suggests a phased classification method
of requirements for a registry by the kind of cancer
and the definitions of items. First, this study classifies items by category, centering around the flows
of the treatment of the patients, and next, classifies
items by division, mainly focusing on the treatment
methods. Finally, items by section are classified
mainly by the treatment history. Items by category
and division are classifications based on the basic
process of the treatment of cancer patients, so they
can be commonly applied.
Result
ŠŠ
Unlike to the category and division commonly applied for registry by cancer types, items by section
showed deferent treatment medthod in the process
of application. So different composition was applied
to them.
Conclusion
ŠŠ
This study suggests phased standard of item classification and if it applys to construct registy, the
study will contribute to building a construction process of the standard regisry by cancer types.
Keywords: Registry, electronic medical records(EMR), cancer, classification
99
Development of Patient Portal for Exchange
Personal Health Record
Ji Hae Kwon, Dong Ouk Kim, Yoo Jin Choi,
Deok Jong Park, Sang Wook Cho
R&D Center, INFINITT Healthcare Co.,Ltd. Seoul,
Republic of Korea
ABSTRACT
Objective
ŠŠ
A paradigm shift in medical services, from being
provider (medical organization)-centric to client (patient)-centric, coupled with aging population that
led to more chronic diseases and higher interest
in health results in the increasing need for collecting, sharing, and exchanging the patient’s personal
health record. This study is aimed at providing a
platform that can meet such need: by implementing
a portal which collects personal health and medical
records and provides them to patients according to
Healthcare standards.
Conclusion
ŠŠ
The outcome of the study achieved Meaningful Use
Stage 2 certification. The developed system has
been implemented and in use in various imaging centers in the U.S. It is our expectation that the portal
can be used for building more convenient and useful
personal health management systems (e.g., ubiquitous healthcare systems) based on further advanced
mobile environments and technologies, when patient
health data collection systems compliant with standards of each medical institution are deployed in full
swing.
Methods
ŠŠ
A portal was designed to collect patients’ personal health records according to the criteria defined
by Meaningful Use regulations. Terms and code
systems for each item were mapped with industry
standards such as SNOMED CT, and were created
in C-CDA, a standard document format, to allow exchange of health records data with different organizations.
Results
ŠŠ
We developed a system that collects and saves patients’ comprehensive medical records (e.g., clinical
studies, imaging studies, and results of examinations) as well as a web portal that displays and prints
the saved health records of patients. The web portal
was designed to enable patients to access in the
mobile environment as well.
Keywords: Personal Health Record, Health Data Exchange, Clinical Terminology, Patient Empowerment
100
2016 Conference of
Asia Pacific Association for Medical Informatics
FHIR for IHE LAW Profile
Sunghwa Lee1, Donghyeong Seong1,
and Byoung-Kee Yi2, 3
1_ Department of Health Sciences and Technology, SAIHST,
Sungkyunkwan University, Seoul, Korea
2_ Dept. of Digital Health, SAIHST, Sungkyunkwan University,
Seoul, Korea
3_ Digital Healthcare Research Center, Samsung Medical Center,
Seoul, Korea
ABSTRACT
With recent advancement in ICT (Information& Communication Technology), an IVD (In-Vitro Diagnosis)
analyzer is expected to be able to interface with
many other devices and services such as smart
phones, tablets, Cloud services, and so on, which
puts a great burden on product development. Traditionally, an interfacing capability of an IVD analyzer
was limited to LIS (Lab Information Systems), for
which IHE published the LAW profile based on HL7
V2 message standard, where most of the data that
can be generated by an IVD analyzer are defined.
While the interfacing technology based on HL7 V2
works with LIS, it is not well suited to mobile devices
and Cloud services.
HL7 FHIR (Fast Healthcare Interoperability Resources) is a new standard candidate based on upto-date web standards such as REST and OAUTH
that is being developed in a way that all existing HL7
standards including V2, V3, and CDA, and focuses
on developers for easy implementation. In this study,
we redefine IHE LAW profile in terms of HL7 FHIR
by converting V2 messages to FHIR and apply the
new profile to an IVD analyzer so that it can easily
interface with emerging devices and services as well
as the existing LIS.
Objectives
ŠŠ
We redefine IHE LAW profile in terms of HL7 FHIR.
Methods
ŠŠ
1. Review the HL7 V2 messages in IHE LAW Profile
The HL7 V2 messages in LAW Profile use 16
segments which are not the same as defined in
HL7 V2. Several segments restrict some fields
101
and their associated dataset according to the
use cases defined in LAW Profile, which we must
take into consideration.
2. Select FHIR base resources
HL7 V2 message segments and FHIR base resources are defined and categorized according
to their usage. For example, ORC segment of
HL7 V2 and Order resource of FHIR are defined
to contain or explain order information. We select
the base resources of FHIR from this point of
view. Also we check the mapping table between
HL7 V2 segments and FHIR resources on the
FHIR official website before the selection process.
3. FHIR resource profiling
There may be some pieces of information in LAW
V2 messages that cannot be expressed by FHIR
base resources. In such a case, we create additional elements as "extension" to appropriate
base resources and also put restrictions on the
elements in base resources that will not be used.
4. FHIR messaging
A conventional transaction in LAW profile is
based on transmission of information from the
actor that has information. We follow the guidance for messaging in the FHIR website. All of
the resources for a transaction are wrapped in a
Bundle resource and the type element value of
this Bundle resource should be set as "message".
Moreover this Bundle resource should contain
MessageHeader resource as first entry resource.
Both of the two actors in the LAW Profile should
perform a server and a client role illustrated below.
Results
ŠŠ
We were able to redefine 17 of HL7 V2 segments
in the LAW profile using 10 FHIR resources, namely,
MessageHeader, OperationDefinition, Specimen,
Patient, Encounter, Bundle, Order, OrderResponse,
Observation, and Substance. Further, we defined a
total of 34 extension elements for Specimen, Order, OrderResponse, and Observation resources.
For verification of these conversion, we developed
an emulator for an implementation of LAW profile
for Samsung PT10 device that had passed an IHE
Connectathon. Under various testing scenarios, it
worked well and, hence, partly verified the correctness of our implementation.
Conclusion
ŠŠ
We converted HL7 V2-based transactions in IHE
LAW profile to ones based on FHIR. We also verified
the correctness of conversion using a simulator.
Although a more rigorous verification such as Connectathon-style testing must be required, this study
provides an ample evidence of the effectiveness of
FHIR-based implementations for interface development of IVD analyzers.
Keywords: Standards, FHIR, IHE LAW profile, IVD
102
2016 Conference of
Asia Pacific Association for Medical Informatics
Design of middleware solution for the vendor neutral
archive system
Min Soo Park, Dong Ouk Kim, Yong Seok Kim
R&D Center, INFINITT Healthcare Co.,Ltd. Seoul,
ABSTRACT
Objective
ŠŠ
Vendor Neutral Archive (VNA) solutions are increasingly being adopted for integrated management of
various medical images and contents. The introduction of a VNA requires integration across medical
information systems based on contents, rather than
mere physical integration. This necessitates an efficient data relay process. The aim of this study is to
design a middleware that classifies main relay processing functions needed for VNA and that enables
processing of large data contents.
Methods
ŠŠ
A middleware was designed for routing and
prefetching, among main functions of VNA: A Message-Oriented Middleware (MOM) was designed to
which Advanced Message Queuing Protocol (AMQP)
is applicable, instead of using a database in a queuebased system.
Keywords: PACS, VNA, Middleware, DICOM, MOM
103
Results
ŠŠ
Construction of a middleware with a message broker resulted in the increased total throughput and a
system that enables quantitative performance monitoring when using routing and prefetching.
Conclusion
ŠŠ
For a VNA to function as a long-term archive for
large-scale data, the delay in data processing should
be minimized and the total throughput should be
maximized, by carefully designing a back-end architecture. A message-based middleware used for such
an architecture can increase the total throughput as
well as allow a high-availability, fault-tolerant structure.
Design and Implementation of SMART guide system based on
integrated data modeling
Lim Young Hee, Kim Ji Hae,
Hyung Il Lee
Division of Medical information and Technology
(YONSEI UNIVERSITY HEALTH SYSTEM,
SEOUL, KOREA)
ABSTRACT
Objective
ŠŠ
The purpose of this study is to design, implementation, and evaluation of SMART guide system that
utilizes PDP/LCDs and a hospital information system(HIS) to address the difficulties and insufficiency
that outpatients meet in finding hospital facilities,
recognizing each examination’s guide and schedule,
and accessing personalized medical and administrative information. Those should can manage independent control based on each examination process
and subject without any constraint.
Methods
ŠŠ
The present study was conducted in a fully digitized
tertiary university hospital in South Korea. We developed outpatient SMART guide system that consists
of data controller to manage to present display contents and time on receiver, it can be also covered
about cycle, end lag time, waiting time, guide script
etc., and a kind of display device, LCD/PDP based
guide application, a guide server which includes the
integrated master based on patient flow and data
process in every patient examination, and interfaces
with the HIS. Applying UX user interface, End users
who play role on master manager easily without IT
staffs.
Results
ŠŠ
We confirmed that SMART guide system can be
applied effectively to visiting outpatient who have
an examination. We completed to apply about 300
examinations, also it includes each process and
events. After doing a pilot test, we implemented
1,500 LCD/PDP which related to examinations at
the same time. The satisfaction rating of the system
was high, showing its utility as a patient-centered
hospital service.
Conclusion
ŠŠ
The design of SMART guide system offers reduction
delay time of patient medical treatment, obtains
higher score in customer service satisfaction, and
improve response time effectively according to medical condition and procedure. There is the improvement of hospital information system in ubiquitous
environment.
Keywords: SMART guide system, HIS (Hospital information system), ubiquitous environment, Outpatient process, a patient-centered hospital service.
104
2016 Conference of
Asia Pacific Association for Medical Informatics
Design and Development of a Medical Information
Support System
YANG Peixiang*, ZHAO Dongsheng,
WANG Xiaolei, ZHANG Wei
Institute of Health Service and Medical Information, Academy of Military Medical Sciences
Beijing, China
ABSTRACT
With the development and application of Internet,
information research institutes are confronted with
both opportunities and challenges. Establishing the
information support system based on Web has become a trend.
Objectives
ŠŠ
To develop a medical information support system
based on the needs of medical information research
institute, to realize the comprehensive integration,
effective utilization and knowledge service of medical information resources.
Methods
ŠŠ
Based on the information classification principle,
the medical information accumulated by the information research institute has been divided into journal articles, research reports, books, dissertations,
conference information, network news, laws and
regulations, standards, experts and scholars, organizations, professional vocabulary and other 12 kinds
of types. Using metadata storage, full-text retrieval,
data mining, web crawler technology, the medical
information support system has been established
by adopting the J2EE technology based on the
Browser/Server architecture. Windows 2008 Server
(64 bits), JDK1.7, Tomcat7.0 and MySql5.5 were
employed as the developmental platform to realize
the system functions. Users can access the system
through the browser, using a dedicated account.
Results
ŠŠ
The medical information support system can exercise overall management. Internet information capture. Periodical information collection, classification,
storage and release of the main sites of Internet bio-
logical medicine. Resource processing and indexing.
The processing of medical information resources,
indexing, storage and other operations. Information
storage management. Add, deletion, and modification the information stored in the database. Subject
classified navigation. Organization, navigation and
browse the information stored in the database according to the set of subjects. Information query and
retrieval. The information stored in the database can
be general query and full-text retrieval. Visual mining
analysis. Construction of medical information network, based on keywords, experts, relevant literature
to provide knowledge discovery. Statistical analysis
of data. Through multi-dimensional statistical analysis, the sources and utilization of the information can
be analyzed and displayed in a variety of graphics. Integrated management of the system. Including data
synchronization, interaction, backup and recovery,
also includes security authentication mechanism for
information dissemination to provide security guarantees.
Conclusion
ŠŠ
We have successfully put the medical information
support system into use since April 2014, after 2
years of operation, it has accumulated nearly 10
thousand data records. The system is characterized by the friendly interface, diversified retrieval
methods, rapid response and complete output. As
an attempt of the information research institute, the
medical information support system significantly
promoted the institute’s information construction,
improved the intelligence research quality and efficiency, and supported the academy’s innovation and
decision-making in the area of medicine.
Keywords: Medical Informatics; Information Retrieval; Information System
105
P2
November 4 (Fri), 2016
Poster
Session
•
Exploring factors affecting the time to fall from admission using EHR data
•
The development of eHR domain content in HKSAR
•
A Fast Microbial Detection Algorithm based on High-Throughput Sequencing Data
•
Patient Engagement Function Usage Trends in a Hospital-tethered Mobile Personal Health Records
•
User requirement analysis for hyperlipidemia management application
•
Identification of Strategies to Promote Positive Mental Health among the Adolescents using
a Mobile App
•
A Design of Smartphone App for Children’s Weight Control with Their Parents
•
Prescribing Error Reducing After the Implementation of Computerized Physician Order Entry
(CPOE) in Thai University Hospital
•
Computerized Physician Order Entry(CPOE) Improved Financial Outcome in Thai University Hospital
•
An Investigation Report of Citizen's attitudes toward the handling of Electronical Medical and
Medicine Information.
•
A Research Study on the Design of Secure HIS that Based on International Standards for General
Hospitals : with a focus on ISO 17090, ISO 22600 and ISO 27799
•
Non-smoking Webtoon opinion mining, a non-smoking awareness through the Analysis of posting
comments
•
Childhood Vaccination Ontology for Social Data Analysis
•
Development of Medical Informatics Principles for Data Repository in Different Healthcare IT
Platforms in Singapore
•
The mediating effect of depression on sleep disorder and fatigue among reproductive aged
women with Psoriasis
•
Reliability and Validity of the Persian Version of the Decisional Conflict Scale in Selecting Mode of
Child Delivery
•
Using Simulation to assess user needs for the Point of Care (POC) Robot development
106
2016 Conference of
Asia Pacific Association for Medical Informatics
Exploring factors affecting the time to fall from admission using
EHR data
Hyesil Jung and Hyeoun-Ae Park
College of Nursing, Seoul National University,
Seoul, South Korea
ABSTRACT
Objective
ŠŠ
The aim of this study was to explore factors affecting length of fall free days from admission using
electronic health record (EHR) data.
Methods
ŠŠ
This was a retrospective observational study of 71
fallers who were admitted to neurology, neurosurgery, hematology and medical oncology wards in a
tertiary-care university hospital during the period
from January 2015 to May 2016. A faller is defined
as a patient with nursing narratives such as ‘Fall’,
‘Slip and Fall’, ‘Fall off the bed’, ‘Fling oneself to the
floor’, ‘Fall down during a move’ and ‘Slip down during
a move’ in nursing notes. The multiple regression
and cox’s proportional hazards regression analyses
were used to identify factors affecting a time period
between admission and fall event. Sex, age, Hendrich-fall risk score and patient severity score were
included as independent variables in analyses.
Results
ŠŠ
There were 48 male (67.6%) and 23 female (32.4%)
fallers with a mean age 64.8 and 64.9 years, respectively. The average Hendrich-fall risk score and patient severity score at admission were 5.6 and 24.0,
respectively. The average time to fall from admission
was 17.4 days. Figure 1 shows the Kaplan-Meier
survival curve depicting the time to fall from admission of the 71 fallers. The multiple regression model
(p=.0483) and Cox’s proportional hazards regression
model (p=.0292) were statistically significant.
Table 1 shows the estimates of regression coefficients for the two models. Hendrich-fall risk score
and patient severity score were statistically significant variables to predict the length of time between
admission and fall event in the models.
The higher Hendrich-fall risk score is, the earlier patient fell and the higher patient severity score, the
later patient fell.
Even though sex and age were not statistically significant, the older and male tend to fall earlier than
the younger and female patients.
Conclusion
ŠŠ
This study is a pilot study to explore the relationship
the time to fall from admission with four key factors
of fall. We will extend the study by including control
groups and more risk factors to develop a fall prediction model.
Figure 1 Kaplan-Meier survival curve of falls
Table1 Results of two analyses
Parameter
Sex0(female)
Age
Fall-risk score
at admission
Severity
score at
admission
Multiple regression analysis
Cox’s PH regression analysis
Cox’s PH regression analysis
SE
Pr>Chisq
0.4525
0.1210
Parameter Estimate
-0.21570
0.00994
0.27995
0.00941
0.4410
0.2912
Hazard
Ratio
0.806
1.010
0.84149623
0.0330
0.15509
0.06155
0.0118
1.168
0.31834254
0.0631
-0.05719
0.02470
0.0206
0.944
Estimate
SE
Pr>t
3.35389367
-0.24585262
4.43748356
0.15643199
-1.83359790
0.60216930
Keywords: fall, EHR data, nursing narratives, survivor analysis
107
The development of eHR domain content in HKSAR
Tsui HH1,2, Szeto WHK1,2, Fung HV1,2, Lau PKM1, 2,
Hung SYV1, 2, Wong SHJ1, 3, Mok HMJ1, 2 and Cheung NT1
ABSTRACT
The Electronic Health Record Sharing System
(eHRSS) is a territory-wide platform that allows sharing of electronic womb-to-tomb health records between public and private healthcare providers (HCP)
with healthcare recipients’ consent. eHRSS was
launched on March 13 2016, with the aims to improve the quality of healthcare delivery by enhancing
the continuity of care, improving the integration of different healthcare services for individual patients, and
promoting the public-private collaboration.
Objectives
ŠŠ
This paper aimed to describe the process of developing the data content of the first stage eHRSS.
Methods
ŠŠ
The stage 1 eHR sharable scope and the data content of each domain were developed by undergoing
the following journey:
The initial set of eHR content standards were defined with reference to local or international standards
e.g. ASTM E1384 Content & Structure of Electronic
Health Record and E369 Continuity of Care Record
(CCR), HL7 standards, and Hospital Authority’s (HA)
data for electronic patient record (ePR). Multi-level
standards compliance strategy was formulated to cater for different levels of computer adoption at various
HCPs; i.e. Level 1, free-text or PDF document; Level
2, structured data with local value and Level 3, structured data with standard value.
Domain groups (DG) were set up to oversee the
standards of data sharing for specific domains, such
as, drug, laboratory. Consultation was sought from
expert advisory group for domains that have no DG.
Meanwhile, gap analysis with HA, Clinical Management
System (CMS) On-ramp and CMS Adaptation Modules was performed to ensure interoperability. Briefing
sessions were held and comment was sought from local health care providers on the data standards. Also,
the display format of the eHR data content in the eHR
Viewer was proposed with ongoing enhancement.
Finally, the data content and information standards
were endorsed by Coordinating Group on eHR Con-
1_ eHR Information Standards Office, Hong Kong
2_ Information Technology and Health Informatics Division, Head
Office, Hospital Authority, HKSAR
3_ Chief Pharmacist’s Office, Hospital Authority, HKSAR
tent & Information Standards. The related standards
documents were published to http://www.ehealth.gov.
hk for reference.
The defined data standards provided a framework of
data to be shared in eHRSS. It facilitated the testing
of compliance standards and ensured data quality uploaded to eHRSS.
Results
ŠŠ
There were 15 domains data standards developed
in the phase 1 eHRSS (figure 1). 1,458 records covered with 10 domains were successfully uploaded
in the first week of eHRSS launched. As more and
more participant joined eHRSS, the number of uploaded records was increased to 13,011,152 which
covered with all the 15 domains in the 4th week of
eHRSS launched.
Conclusion
ŠŠ
eHRSS enables two-way sharing of eHR contributed
by both public and private sectors. The eHR Information Standards Office will review and update the sharable scope as appropriate and more domains standards will be developed in the later stage of eHRSS.
Figure 1 Phase 1 eHR content
Domain
Healthcare recipient index
Encounter
Referral
Clinical note/summary
Adverse drug reaction
Allergy
Problem
Procedure
Birth Record
Dispensing record
Prescribing record
Immunization
Laboratory
Radiology examination
Others investigations
Level of compliance
Level 1 Level 2 Level 3
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
Keywords: eHealth Record, EHR content, eHRSS
108
2016 Conference of
Asia Pacific Association for Medical Informatics
A Fast Microbial Detection Algorithm based on
High-Throughput Sequencing Data
LI Jiangyu, MAO Yiqing, WANG Xiaolei,
ZHAO Dongsheng
Institute of Health Service and Medical Information, Academy of
Military Medical Sciences Beijing, China
ABSTRACT
Objective
ŠŠ
According to the ‘output while sequencing’ characteristic, design a localized rapid microbial detection
algorithm. The algorithm analyzes the sequencing
data while sequencing, which can improve the speed
of pathogenic microbial detection.
Method
ŠŠ
A ‘analysis while sequencing’ method is used to analyze the sequencing data while sequencing, the
method is using the detection algorithm to analyze
the sequencing data generated by the sequencer
periodically. In each analysis, the reference microbial
genomes are grouped. For each group the algorithm
extracts sequencing reads mapped well to the microbial genomes and then filters the human genome
data in the reads extracted, then the algorithm assembles the reads left and aligns the assembled
contigs to the microbial genomes. The algorithm
flow ends if the microbial species are detected from
the comparison results. Otherwise the algorithm
continues with the next group of reference microbial genomes. If the algorithm still cannot find out
the microbial species when all reference microbial
genomes are used, then it filters out the human sequencing reads in the rest of the sequencing data
and assembles the reads left. If the assembled
contigs cannot map to any reference microbial genomes, these contigs are detected as genome segments of unknown pathogenic microorganisms.
Result
ŠŠ
First, the new detection algorithm was used to analyze the simulated data and real sequencing data respectively, and RINS is used for comparison. For the
simulated data, the new algorithm achieves speedup
of 10 compared with RINS when the microorganisms in the sample have ref-genomes, the analysis
results are consistently and the new algorithm gets
longer contigs. The new algorithm achieves an average speedup of 9 when the microorganisms in
the sample don’t have ref-genomes, and it obtains a
more complete original sequence. For the real data,
the new algorithm achieves an average speedup of
3 compared with RINS, and the detection results
of the two algorithms are the same. When verifying
the Sequencing-by-side method, a reliable result
can be obtained with a certain scale of sequencing
data. Along with the increase of sequencing data,
the number of contigs obtained by the algorithm is
reducing and the length of contigs is increasing.
Conclusion
ŠŠ
The ‘analysis while sequencing’ methods implemented in this paper can improve pathogen detection
speed, and have good application prospects.
Keywords: high throughput sequencing; microorganism detecting; sequencing assembly; algorithm
109
Patient Engagement Function Usage Trends in
a Hospital-tethered Mobile Personal Health Records
Jae-Ho Lee1,2, Ji-Young Kim3, Jeong-Hoon Kim3, ByungWook Sung3, Yurang Park1, and Woo Sung Kim1,3,4
1_ Department of Biomedical Informatics, Asan Medical Center, Seoul, Korea
2_ Department of Emergency Medicine, University of Ulsan College of Medicine,
Seoul, Korea
3_ Medical Information Office, Asan Medical Center, Seoul, Korea
4_ Department of Pulmonary & Critical Care Medicine, University of Ulsan
College of Medicine, Seoul, Korea
ABSTRACT
Objective
ŠŠ
Patient engagement functions (PEFs) is an important factor of continuous use of mobile personal record (mPHRs). A hospital-tethered mPHR which was
actively used in Korea since December, 2011 was
recently updated to provide more PEFs to patients.
By analyzing usage trends of PEFs, we tried to find
which functions were used frequently or seldom to
guide next upgrade of the mPHR.
Methods
ŠŠ
PEFs were defined as that of patients’ active action
needed for their health and disease management.
Functions of health data management entry, medication entry, allergy entry, phone call, diary entry for
three diseases, my health summary download, and
questionnaire entry for cancers and inflammatory
bowel diseases were included. Phone call, my health
summary download, diary entry, questionnaire entry
were newly added functions to the upgraded mPHR.
Six months’ log data of PEFs and log data of sign-up,
users, and access to the mPHR were gathered and
tendency of their usage were analyzed.
PEF (total: 10870, daily mean: 59.7) then followed
by Diary entry (2270, 12.5), Medication entry (1774,
9.7), My health summary download (1262, 6.9),
Questionnaires (1248, 6.9), Phone call (566, 3.1),
and Allergy entry (43, 0.2). Within the health data
entry, blood sugar test was the most used (5246,
28.8), then flowered by Blood pressure and heart
rate (3466, 19.6), Body mass index (1503, 8.3), 10
years’ Cardiovascular risk (354, 1.9), and Metabolic
syndrome (301, 1.7). Except Allergy entry menu, all
PEFs usages were increased as time passed. But
usages of only three PEFs—Questionnaires, Dairy
entry, and My health summary download– were increased when divided by users logged-in.
Conclusion
ŠŠ
The mPHR usage was increased after upgrade. The
daily usages of three of four newly added PEF functions were increased but those of previously existed
PEFs and Phone call were slightly decreased. More
in-depth studies of PEFs usage trends are needed
for improving patient engagement including previously existed PEFs usage trends.
Results
ŠŠ
During the study periods, total sign-up, users, and
access were 20947, 169621, and 336113 (115.1/
day, 932.0/day, 1846.8/day) and they showed increase in use. Health data entry was the most used
Keywords: Personal health record; Mobile health; Patient engagement; Patient centeredness
110
2016 Conference of
Asia Pacific Association for Medical Informatics
User requirement analysis for
hyperlipidemia management application
Joo Yun Lee, Hyeoun-Ae Park
College of Nursing, Seoul National University,
Seoul, South Korea
ABSTRACT
Objective
ŠŠ
The objectives of this study are to survey user requirement for tailored information and feedback
from the app on self-monitoring of diet, exercise,
smoking cessation, medication and weight control
for hyperlipidemia management app.
Methods
ŠŠ
Paper-based survey was conducted on 30 patients
who were diagnosed as hyperlipidemia. Questionnaires contained 7 items on demographic information, 10 items on tailored information, and 22 items
on self-monitoring with feedback. We asked patients
to rate how much they think they need tailored information on the importance of and how to manage
each life style. For self-monitoring, we asked patients
to rate how much they think they need displaying daily activity, displaying activity trend, reminder for data
input, reminder for lifestyle activity as feedback from
the app, and goal setting function using 5-point rating
scale (from 1 = the least to 5 = the most). We allowed
patients to give any suggestion with open-ended
question. Rating scores were analyzed by descriptive
statistics using means and SDs.
Results
ŠŠ
Among 30 patients, 22 patients were male and 8
were female. Their mean age was 52.87 years (Min.
29 years, Max. 70 years). All patients were smartphone users. For tailored information, the patients
rated information on the importance and management methods of life style component higher than
4.0 except for the information on smoking cessation
management method. For self-monitoring, they rated goal setting function relatively high (3.90 – 4.43)
and reminding data input as a feedback on self-monitoring relatively low (3.57-4.07) (Table 1). Three patient suggested a function to compute the likelihood
of developing hyperlipidemia complication.
Conclusion
ŠŠ
Most patients think they need tailored information
for lifestyle management. Highly rated functions
were goal setting. However, patients do not think
they need remind alarms for their behaviors. We will
develop hyperlipidemia management app reflecting
user requirements.
Table1 Rating score for tailored information and self-monitoring of hyperlipidemia management app
Tailored information
Feedback on self-monitoring
Importance
Management
Method
Displaying
daily activity
Displaying
trend
Reminding
data input
Reminding
activity
Goal
setting
Diet
4.47
4.50
4.33
4.23
3.77
-
4.33
Exercise
4.53
4.67
4.30
4.23
3.87
3.80
4.43
Smoking
4.13
3.83
4.03
3.80
3.57
-
3.90
Medication
4.10
4.03
3.93
4.01
4.00
3.83
4.20
Weight control
4.40
4.43
-
4.37
4.07
-
4.37
Keywords: Mobile health, Hypercholesterolemia, Health behavior, Needs Assessment
111
Identification of Strategies to Promote Positive Mental Health
among the Adolescents using a Mobile App
Ji-Hyun Lee , Hyeoun-Ae Park
College of Nursing, Seoul National University,
Seoul, South Korea
ABSTRACT
Objective
ŠŠ
The aim of this study is to identify strategies for a
mobile app to promote positive mental health in the
adolescents.
Methods
ŠŠ
First, social emotional learning competencies as a
framework of educational contents to promote the
positive mental health in the adolescents were identified by reviewing the clinical practice guidelines
published by the NICE and the CASEL [1], and related literature. Second, skill sets to develop theses
competencies were identified by reviewing literatures and validated by two mental health nurse-specialists. Third, strategies to realize these skills were
identified by reviewing the literatures on the mental
health apps and studying functions of the mental
health apps. The appropriateness and preference of
these strategies were surveyed with three school
health teachers and 13 middle school students. Finally, competencies, skill sets and strategies were
linked and the linkage was validated with three nursing informatics experts.
Results
ŠŠ
Table 1 displays five competencies, eight skill sets,
and five strategies identified for the app and their
relationships. Two mental health nurse-specialists
agreed eight skill sets are enough to achieve the
goal of the app. School health nurses and students
answered five strategies are appropriate to develop
eight skills. School health nurses preferred self-monitoring and tailored education strategies. However,
students preferred gamification and social media
strategies. Finally, nursing informatics approved the
strategies to be implemented on the app to achieve
the skill sets identified for the competencies required for positive mental health promotions in adolescents.
Conclusion
ŠŠ
A mobile mental health app to promote positive
mental health in adolescences will be developed
based on the strategies identified in this study
This work was supported by the National Research
Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2010-0028631).
Table1 Competencies, skill sets and strategies for a mobile mental health app
Competencies
Self-awareness
Self- management
Social awareness
Skills sets
Strategies
Selfmonitoring
Tailored
education
Social
media
Gamification
Reminder
O
X
X
O
O
O
X
X
O
X
Managing stress
X
O
X
X
X
Mobilizing positive motivation
X
O
X
X
X
Recognizing others’ emotions
X
X
O
X
X
Recognizing and naming one’s
emotions
Understanding the reasons and
circumstances for feeling
Appreciating diversity
X
X
O
X
X
Relationship skills
Help-seeking
X
O
X
X
X
Responsible
decision making
Engaging in self-reflection
O
X
X
O
X
Keywords: Mental health, Health Promotion, Mobile Apps, Adolescents
112
2016 Conference of
Asia Pacific Association for Medical Informatics
A Design of Smartphone App for Children’s Weight Control
with Their Parents
Ahjung Byun, RN1, Meihua Piao, RN, M.Sc1, Jisan
Lee, RN, M.Sc3, Jeongeun Kim, RN, PhD, INS1, 2, 3,
Hyeongsuk Lee, RN, MSN1,4
1_ College of Nursing, Seoul National University, Seoul, Republic of Korea
2_ Research Institute of Nursing Science, Seoul National University, Seoul,
Republic of Korea
3_ Interdisciplinary Program of Medical Informatics, Seoul National University,
Seoul, Republic of Korea
4_ Seoul Women’s College of Nursing, Seoul, Republic of Korea
ABSTRACT
Objective
ŠŠ
Pediatric obesity concerns healthcare professionals
around the world due to its increasing harmful effects.
Overweight or obese children are prone to various
chronic diseases and also influenced psychologically
and socially [1]. According to previous studies, parents had positive influence on children when cared
well with proper diet and physical activity [2]. Along
with lack of will, lives of elementary school students
in Korea are centered around school work and private
education that it is difficult for them to control weight
and stay healthy [3]. Currently, smartphone application (app) has been emerged as an alternative means
to healthcare industry and it is possible to apply personalized intervention to obese children in a cost-effective way. In this study, we depicted a smartphone
app design so as to actually realize it.
Methods
ŠŠ
In order to design an app, we used the LUCID framework [4]. It helps the developers to create the product
they envisioned. Its first step is to envision. We wanted
to create an app that helps elementary school students to control their weight with their parents’ guiding and stay healthy by practicing good acts of staying
healthy. The second step is Discovery Stage. So as to
find out the users’ needs, we conducted focus group
interview and a survey. Based on the results, the third
step, design foundation, has been done.
Results
ŠŠ
Figure 1 shows the prototype of the app design
based on the results of focus group interview and
survey. The content itself contains the ideas about
habit loop by Charles Duhigg [5]. This habit loop
consists of three steps: cue, routine, and reward. By
applying this means, it helps children to build a habit
of practicing healthy lifestyle.
Conclusion
ŠŠ
An app design for children’s weight control with their
parents has been depicted and this would be applied to real users after implementation. With further
study, the effect of this app would be confirmed and
this could be used as a new method for elementary
school students’ new means to control their weight
with their parents.
Figure 1 The prototype of the app design
Keywords: Pediatric Obesity, Smartphone App Design, Parents, Elementary School, Weight Control
113
Prescribing Error Reducing After the Implementation of
Computerized Physician Order Entry (CPOE) in
Thai University Hospital
Soontornpas R, Jiravuttipong A,
Pongskul C, Bhoomibunchoo C
Faculty of Medicine, Khon Kaen University, Khon
Kaen, THAILAND
ABSTRACT
Introduction
ŠŠ
Srinagarind Hospital, a 1000 beds university hospital had implemented new Hospital Information
System (HIS) since December 2014. New Computerized Physician Order Entry (CPOE) is a hospital
management and information system for preventing
or minimizing unwanted events that may cause harm
to the patients called medication errors. This system
may decrease prescribing error and improve patient
safety since physicians can send medication order
directly to the pharmacy service department instead
of using paper-base prescription with hand writing.
Moreover, the commercial MIMS drug information
was incorporated to the CPOE and the physician
can select the medication in hospital formulary from
drop down list in the software; meanwhile, drug-drug
interaction, drug duplication and drug allergic history
can be displayed and allow the physician to make
final decision. CPOE has been a major institutional
change at Srinagarind Hospital to minimize the prescribing errors and may enhance patient safety and
efficiency of the treatment. As the system has been
implemented since December 2014, therefore, the
monitoring of reported medication error and the
constructive feedback should be performed to assess the effectiveness of this strategy.
Objectives
ŠŠ
To explore the outcome of CPOE implementation on
the incidence of prescribing error.
Methods
ŠŠ
This retrospective descriptive study was performed
at Srinagarind Hospital, a 1000-bed teaching hos-
pital, Khon Kaen University, Thailand in July 2016.
The data about prescribing errors were obtained
from the hospital’s voluntary paper-based reporting
system for medication errors which were already
assessed by pharmacist in pharmacy service department before CPOE implementation (2014) and
after CPOE implementation (2015). The report included incident date and time, report date, location,
medical service, type of error, medication involved,
severity and the name of person who notified and/
or reported. The incidence of medication error was
calculated as the number of event divided by 1000
prescriptions and used as service indicator. Then,
the outcome of COPE was assessed by comparing
these service indicators between pre and post implementation.
Results
ŠŠ
Prior to CPOE implementation, the incidences of
prescribing error were ranged between 5.14-6.50
and 3.35-12.98 events/1000 prescriptions in outpatient and inpatients, respectively. A year after the
CPOE implementation, the incidence of prescribing
error was decreased to 0.22 and 0.30 events/1000
prescriptions in outpatients and inpatients, respectively, and meant that the incidences of prescribing
errors were reduced approximately 99.9% in both
outpatient and inpatient.
Conclusion
ŠŠ
CPOE system could reduce prescription error and
may improve the patient safety. Undoubtedly patient
safety is the core benefit of CPOE implementation.
Keywords: CPOE, Prescribing error, University hospital
114
2016 Conference of
Asia Pacific Association for Medical Informatics
Computerized Physician Order Entry (CPOE) Improved Financial
Outcome in Thai University Hospital
Pongskul C, Jiravuttipong A, Manokhun D,
Poanariang K, Bhoomibunchoo C, Soontornpas R
Faculty of Medicine, Khon Kaen University, Khon
Kaen, THAILAND
ABSTRACT
Introduction
ŠŠ
Thailand has three major health insurance schemes
that together provide access to health coverage for
all of its citizens. The Civil Servant Medical Benefit
Scheme (CSMBS) provides health care to government employees. The Social Security Scheme (SSS)
is a compulsory insurance program for employees of
private businesses. The largest insurance program
is the Universal Coverage (UC) scheme. In the past,
paper prescriptions by doctors were entered into
the pharmacy information system by a pharmacist.
This human-dependent process unavoidably led to
prescription errors and reimbursement errors. The
former increased patient risks while the latter initiated financial risks. Srinagarind Hospital, a 1000
beds university hospital implemented new Hospital
Information System (HIS) since December 2014.
New Computerized Physician Order Entry (CPOE)
is designed to reduce both types of errors. Doctors
can order through CPOE, with embedded Clinical
Decision Support System. CPOE will automatically
allocate medications by primary health insurance
and payer agreement. Data set up is based on drug
type, encounter, agreement type, and payer’s office.
Each patient can be entitled to medicines from
various insurance schemes per one visit. After verification by financial office, data will be populated and
sent out to the claim office of the applicable payers.
That information will be validated by payer’s format
and rules and sent back to the hospital for correction. Once verified, the reimbursement process will
complete.
Objectives
ŠŠ
To explore the financial benefits of CPOE in reducing
claim denial and increasing reimbursement amount
in CSMBS scheme.
Keywords: CPOE, University hospital, Financial, Reimbursement
115
Methods
ŠŠ
Outpatient reimbursement data of 24 months before and 12 months after the new CPOE implementation were analyzed. Claims denied and error data
were collected from file acknowledgment by the
payers. The claim denial is identified by the incomplete claim data, i.e. patient information, encounter
information, product category, drug code, compatibility of drug code to hospital drug catalogue,
reason for using non-essential drug, and prescriber
information. Percent of claim denial is defined by the
ratio of number of denied claims to the number of
total claim transactions multiply by 100. Two financial outcomes to be observed are percent of change
in annual reimbursement amount and percent of
change in annual reimbursement amount per visit.
Results: Prior to CPOE implementation, 27.5 percent of reimbursements in CSMBS scheme were
rejected. While some can be claimed after several
corrections, there were many reimbursements left
unclaimed. A year after the CPOE implementation,
those figures reduced to 4.57 percent. Annual reimbursement amount increased by 40.5 percent while
annual reimbursement amount per visit increased by
8.5 percent.
Conclusion:
ŠŠ
Undoubtedly patient safety is the core benefit of
CPOE implementation, however, it also evidently
helped decrease claiming errors and significantly
improved financial outcome. Data set up was very
important in preventing such errors and led to zero
claim denial.
An Investigation Report of Citizen's attitudes toward the handling
of Electronical Medical and Medicine Information.
Mayumi Yoshida1, Katsuya Tanaka2,
Ryuichi Yamamoto1
1_ Medical Information System Development
Center. (Tokyo, JAPAN)
2_ The University of Tokyo Hospital. (Tokyo, JAPAN)
ABSTRACT
Objective
ŠŠ
If adequate PHRs are realized, it will be beneficial
for health management by accumulation and management of lifelong health care information for every
citizen. To control own information is one of basic
concept of privacy, but to control own health information is difficult for general people because of lack of
enough knowledge for medical sciences. So we consider that PHRs should have supportive functions for
privacy protection and data security. For determining
the requirements for adequate PHRs. we planned to
clarify the acceptability of PHRs by ordinary citizen
and also their attitudes toward data security, privacy,
and secondary use for public benefit.
Methods
ŠŠ
We have carried out Web questionnaire 4 times
about the citizen's attitudes toward the handling of
their medical information. The results of the survey
were investigated for age and health conditions of
the respondents, and also analyzed using cross tabulation. We also analyzed results of opinion polls of
users of PHRs already operated in other countries
such as the Blue Button in the U.S., and compared
the results with those of our results.
Results
ŠŠ
We asked about the necessity for PHRs, 82.4% of
the respondents (N=2266) answered "Yes". Those
who answered "No" (N= 399) were then asked why
they think there is no need for PHRs. 66.9% of the
respondents chose "I think we do not need anything
more than the current system that maternity notebooks and medicine notebooks on paper", 65.4%
chose "I am afraid about the high cost", and 59.9%
chose "I am afraid about the security issues such
as information leaks". Regarding the subject that
should services PHRs, the questionnaire allowed
multiple choices, which resulted in "National governments" was chosen most often: 41.8%, then
"Local governments": 32.9%, "Medical institutions":
32.0%, "Private sector enterprises controlled by
national governments": 16.9%, "Private sector enterprises controlled by local governments": 11.5%.
We asked about necessity for cloud service which is
aimed for electronical medicine notebooks for management of drug information by oneself for lifetime
span, 86.9% chose "Yes, if it is free", 4.7% chose
“Yes, even if there is a charge.". So we asked about
usage of smartphone application of standalone digitalized medicine notebooks, “It's being used at present”: 1.9%, “It was used in the past”: 2.5%. Same
tendency was confirmed in case of paper based maternal handbook and electronical one.
Conclusion
ŠŠ
The surveys revealed that most citizens desire adequate PHRs, on the other hand, the survey also
showed that citizens feel vague anxiety about
security issues. Many citizens seem to recognize
the necessity of PHRs and lifetime electronical
medicine notebooks, on the other hand the rate
of usage of electronical medicine notebooks and
electronical maternal handbooks was quite low. We
suppose the paper-based medicine notebook and
the paper-based maternal notebook have already
quite important role in managing medication and
pregnant-delivery-childcare, and citizens notice the
disadvantage of paper based records, but the function of present electronical medicine notebooks and
electronical maternal handbooks is incomplete and
lack of the long time availability, and we conclude it
is still necessary of further exploitation of PHRs and
should be supported by public sectors for achieving
long time availability.
Keywords: Privacy, PHR, Web Questionnaire Survey, Electronical Medicine Notebook, Electronical Maternal Notebook
116
2016 Conference of
Asia Pacific Association for Medical Informatics
A Research Study on the Design of Secure HIS that Based on
International Standards for General Hospitals
: with a focus on ISO 17090, ISO 22600 and ISO 27799
Yong En Lee, Zip Min Chung, Jea Dong Lee,
Hyo Soung Cha and Seung Hyun Chung
Information Technology Team, National Cancer
Center, Goyang, Korea
ABSTRACT
Objective
ŠŠ
Hospitals are able to provide increasingly diverse
services by implementing HIS (Hospital Information
System). But HIS does not based on a information
management standard and therefore has an insufficient level of security at all stages of processes
ranging from data creation to verification. Therefore
a HIS that applies ISO 17090, ISO 22600 and ISO
27799 standards was designed in this research
study to support the implementation of secure
healthcare services.
Methods
ŠŠ
To design a secure HIS, the authentication certificate policies and security requirement specifications based on the PKI (Public Key Infrastructure)
ISO 17090 standard were analyzed. Additional
security requirement specifications were deduced
by analyzing the SSO (Single Sign-On) / EAM (Enterprise Access Management) ISO 22600 standard
recommended for healthcare providers, which were
reflected into the secure HIS designed in this research study.
Results
ŠŠ
It was confirmed to be possible to verify users with
their electronic signatures when storing record archive documents and orders for the long term by
implementing PKI and that the abuse of permissions
by users can be prevented by implementing SSO/
EAM. Also the Interim HIS designed in this research
study that implements a redesigned minimized hardware profile and software stack that is both medical
Keywords: HIS, ISO, PKO, SSO/EAM, healthcare security
117
record and also order oriented, a secure and convenient healthcare service that manages the creation,
storage, utilization and verification processes using
electronic signatures can be provided.
Conclusion
ŠŠ
The Interim HIS designed in this research study applied international standards to the previous information management system that did not based on
any standards. The scope of patient oriented healthcare IT services can be expanded by based on these
standards, which will enable the implementation of
secure system based on increased security protocols.
Non-smoking Webtoon opinion mining, a non-smoking awareness
through the Analysis of posting comments
Dong Kyun Lee1, Seo-Joon Lee2, Jung Doo Kim1,
In Young Choi, Ph.D1
ABSTRACT
1_ Department of Medical Informatics, The Catholic University of Korea,
College of Medicine, Seoul, Korea
2_ BK21PLUS Program in Embodiment: Health-Society Interaction,
Department of Public Health Sciences, Graduate School, Korea University,
Seoul, Korea
Objective
ŠŠ
Smoking in adolescence critically damages adolescents’ physically undeveloped body and also causes genetic damage. Starting to smoke at young age is known
to highly likely to continue throughout adult period. As
a solution, online and offline smoking cessation campaigns are being implemented at national level.
Among those campaigns in 2014, a webtoon called
“Tale of Cigarette” received positive reviews from viewers, publishing a total of 12 episodes in NAVER webtoon. The cartoon amusingly portrayed a friendly character attempting to quit smoking, urging the viewers to
directly participate in smoking cessation campaign and
enlightening them of the benefits of smoking cessation.
Recently, commercial•public brand webtoons have
been developed through a variety of fields’ advertisement platforms, but methods to evaluate their effects
are insufficient. So far, the only way to evaluate the
effect of brand webtoons is limited to ratings, comments, and page views.
Methods
ŠŠ
Therefore, the proposed research analyzed smoking
cessation advertising brand webtoon’s comments
using opinion mining method. The objective was to
analyze the association between frequent morpheme
and morpheme, and to analyze sentiment embedded
in user comments.
For the research objective, all opinion mining related
domestic and international prior researches were
referenced to design preprocessing operation and association analysis to account for collected comments.
Then, analysis results were compared to prior adolescence smoking cessation research results.
Frequent morpheme analysis results showed that
the word ‘Cigarette’ was most frequent, and other
words such as ‘Smoking Cessation’, ‘Human’, ‘Smoker’,
‘Dad’, ‘Please’, ‘Smoking’ followed the ranking order.
Comments were converted through transaction for
association analysis. Support, confidence and lift were
applied to the transaction data to check for association patterns, and results showed high lifts between
the word ‘Cigarette’ and words ‘Smoking’, ‘Tax’, ‘Street’,
‘Smell’, and ‘Smoke’. Sentiment analysis results showed
that negative sentiment in the early episodes changed
into positive sentiment as the episodes progressed.
Conclusion
ŠŠ
In conclusion, in order to successfully advertise smoking cessation through brand webtoon, it was important
to first gain as much viewer attention as possible, and
then to appeal the importance of the webtoon’s story.
Also, it was important to convince the users understand the harmful effects of smoking and to ultimately
make them quit through repeatedly publishing attractive smoking cessation related topics.
Analyzing readers’ opinion through opinion mining of
webtoon comments is envisioned to provide clear indicators when deciding the course of advertisement. Also,
this will enhance the quality of analysis when handling
atypical data such as user comments, and is expected
to be applied in other practical web sites as well.
Results
ŠŠ
A total of 107,264 participants rated the webtoon
“Tale of Cigarette” as 9.91, and preprocessing results
showed that after preprocessing the initial 20,560
comments, a total of 18,117 comments were refined.
Keywords: Smoking Cessation, Webtoon, Comments, Opinion Mining
118
2016 Conference of
Asia Pacific Association for Medical Informatics
Childhood Vaccination Ontology for Social Data Analysis
Jeongah On1, Hyeoun-Ae Park1,2, Tae-min Song3
1_ College of Nursing, Seoul National University, Seoul
2_ Systems Biomedical Informatics Research Center,
Seoul National University, Seoul
3_ Statistics and Information Research Division,
Korea Institute for Health and Social Affairs, Korea
Childhood Vaccination Ontology for Social Data Analysis
ABSTRACT
Jeongah On1, Hyeoun-Ae Park1,2,*, Tae-min Song3
Objective
ŠŠ
Results
ŠŠ
The aim of the study is to develop
childhood
vac- National
In total,
1760 terms
1 Collegea of
Nursing, Seoul
University,
Seoul were extracted. Out of these
cination ontology as a framework for collecting and terms, 638 concepts were extracted. They are con2 Systems
Biomedical
Informatics
Research
University,
cepts Seoul
for 7 National
super classes,
207Seoul
subclasses, 146 atanalyzing social
data to
understand
parents’
sen- Center,
timent, attitude, experience and knowledge about tributes and 278 value sets. The Figure 1 presents
3 Statistics and Information Research Division, Korea Institute for Health and Social Affairs, Korea
childhood vaccination. This ontology will be used to 7 super classes of the childhood vaccination ontolodevelop vaccination policies
and interventions
be- gy
and
the relationship
among the super classes be* Corresponding
Author: Hyeoun-Ae
Park,
E-mail:
[email protected]
fore, during and after vaccination.
fore, during and after vaccination. Personal factors,
structural and sociocultural factor, informational facMethods
tor, and environmental factor affect the vaccination
ŠŠ
ABSTRACT
The ontology was developed using the revised ‘On- intention of the parents. The vaccination intention
tology
Development
101’
We adeterthe vaccination
andforthe
vaccination
Objectives:
The aim of
the methodology.
study is to develop
childhoodaffects
vaccination
ontology as abehavior
framework
collecting
and analyzing
social
to understand
parents’as
sentiment,
attitude, experience
and
knowledgeof
about
childhood
mined
the domain
anddata
scope
of the ontology
fac- behavior
affects the
experience
the parents.
vaccination.
ontology
will be making
used to develop
policies and interventions before, during and after
tors
affectingThis
patients’
decision
processvaccination
for
vaccination.
Conclusion
childhood vaccination ranging from individual factors ŠŠ
to environmental factors. Then we collected terms Childhood vaccination ontology developed in this
Methods: The ontology was developed using the revised ‘Ontology Development 101’ methodology. We
describing
affecting
childhood
study
will bepatients’
used as
a framework
to understand
determinedfactors
the domain
and scope
of the vaccination
ontology as factors
affecting
decision
making process
for
from
national
immunization
guideline
of Korea, factors
liter- toparents’
decision
making
about
vaccinachildhood
vaccination
ranging
from individual
environmental
factors.
Then
we childhood
collected terms
atures
related
vaccination
postings
of parents’
tion and
parents’ sentiment,
experience
describing
factors
affecting and
childhood
vaccination
from national
immunization
guideline ofattitude,
Korea, literatures
related vaccination
andexperience
postings of and
parents’
sentiment, attitude,
experienceabout
and knowledge
childhood
sentiment,
attitude,
knowledge
and knowledge
childhood about
vaccination
before,
vaccination
on social
media. Weon
extracted
terms and
the concepts
hierarchically.
Wedata.
about
childhood
vaccination
socialconcepts
media. from
We theduring,
andclassified
after vaccination
using
social media
defined the
propertiesfrom
and value
sets of the
and theThe
relation
among
the concepts.
extracted
concepts
the terms
andconcept,
classified
result
of the
social media data analysis using
the
concepts
hierarchically.
We
defined
the
properthe
ontology
can
be
for development
of vacciResults: In total, 1760 terms were extracted. Out of these terms, 638 concepts wereused
extracted.
They are concepts
ties
and
value
sets
of
the
concept,
and
the
relation
nation
policies
and
interventions.
for 7 super classes, 207 subclasses, 146 attributes and 278 value sets. The Figure 1 presents 7 super classes of the
among
the concepts.
childhood
vaccination ontology and the relationship among the super classes before, during and after vaccination.
Personal factors, structural and sociocultural factor, informational factor, and environmental factor affect the
vaccination intention of the parents. The vaccination intention affects the vaccination behavior and the vaccination
Figure 1 7 super classes of the childhood vaccination ontology.
behavior affects the experience of the parents.
Figure 1. 7 super classes of the childhood vaccination ontology.
Childhood
vaccination,
Ontology, Social
media
analysis
Keywords:
Conclusion:
Childhood
vaccination
ontology
developed
119
in this study will be used as a framework to understand
parents’ decision making about childhood vaccination and parents’ sentiment, attitude, experience and knowledge
about childhood vaccination before, during, and after vaccination using social media data. The result of the social
media data analysis using the ontology can be used for development of vaccination policies and interventions.
Development of Medical Informatics Principles for Data
Repository in Different Healthcare IT Platforms in Singapore
ABSTRACT
Since the movement into paperless systems in
healthcare management, a myriad of different systems and platforms has emerged. Although most of
the systems boast to have the capability to handle
the requirements of the healthcare institution, different systems are still being used by different healthcare institutions to serve their own purpose.
In Singapore, the healthcare environment in divided
into 6 clusters based on the regional healthcare
systems. These 6 clusters employ different vendors
for their electronic medical records and the national
electronic health record for each Singapore citizen
employs a different system as well. Community
services and other voluntary organisations likewise
have their own platforms to service their needs.
All this different platforms have caused widespread
confusion as to what data should reside in which
systems. It has also been a great challenge to integrate and rationalise all these data to ensure that
there is smooth data transfer to enable continuity of
care and patient safety.
Through various engagements with different stakeholders to understand their workflow, needs and
identifying gaps in the data flow, we attempt to produce a set of informatics guidelines and principles to
allow healthcare providers as well as non-healthcare
providers a clear understanding as to the definition
of different data repository into different IT platforms.
This guideline is based on 6 informatics focus as
documented below.
Maintain the health continuum record with purpose for all providers, through a longitudinal
health record from birth to end of life.
3U-Informatics Focus
Transform the care team clarity, accountability and effectiveness to ensure optimal co-management and co-ownership.
Facilitate joint governance of information/data domains, with the care recipient and patient at
the centre.
Know your population and patient group better to personalize and achieve the best care outcomes.
Support delivery of care and strengthen model of care through facilitating adaptive e-clinical
process flow and personalized e-collaboration.
Support future medicine through ensuring precision informatics is in place and has effective
practical application.
Keywords: aaaaa
120
2016 Conference of
Asia Pacific Association for Medical Informatics
The mediating effect of depression on sleep disorder and fatigue
among reproductive aged women with Psoriasis
Oksoo Kim1,5, Hee Jung Jang2 , Hyangkyu Lee3, Yoon
Hee Kang1, Eunyoung Cho4, and Jiyoung Kim5
1_ College of Nursing, Ewha Womans University,
Seoul, Republic of Korea
2_ Division of Nursing, Hallym University, Chuncheon,
Republic of Korea
3_ College of Nursing, Yonsei University, Seoul, Korea
4_ Department of Dermatology, Warren Alpert Medical
School, Brown University, Providence, RI, USA
5_ Korean Nurses Association, Republic of Korea
ABSTRACT
Objective
ŠŠ
The purpose of the study was to investigate the
mediating effect of sleep disorder in the relationship
between depression and fatigue among Korean reproductive aged women with Psoriasis.
Methods
ŠŠ
The data were collected from Korean Nurse’s Health
Study (KNHS). Female nurses of reproductive age
20 to 45 years old were included. The KNHS was
conducted between March 2013 and December
2015 as a web-based questionnaire survey. The
total number of nurses of reproductive who participated in KNHS Module 1 was 20,922. Of these,
20,463 had no Psoriasis and 495 were clinically diagnosed with Psoriasis; 171 of the latter group were
diagnosed within one year prior to the survey. One
of the main variables, fatigue, was measured using
Chalder et al (1993)’s questionnaires. Depression
was measured with the Patient Health Questionnaire
(PHQ) 9, a self-reported, brief depression assessment tool developed to determine depression levels
based on the DMS-IV. In order to measure the sleep
disorder degree, Jenkins Sleep Questionnaire (Jenkins, Stanton, Niemcryk, & Rose, 1988) were used.
Data were analyzed a hierarchical-multiple regression analysis as suggested by Baron & Kenny with
Sobel test for mediating effect, and were analyzed
t-test, ANOVA, Pearson correlation coefficients according to statistical purpose.
Keywords: Psoriasis, Depression, Fatigue, Sleep, Women
121
Results
ŠŠ
There were differences in depression, sleep disorder according to age, marital status, shift work and
other variables had no significant differences. First
step, There were correlation between depression,
sleep disorder, and fatigue. There was a positive relationship between sleep disorder and depression
(β=.663, p<.000) and Second, a positive relationship between sleep disorder and fatigue (β=.484,
p<.000). Third step, a positive relationship between
sleep disorder, depression, and fatigue. So, there
was a partial mediating effect of sleep disorder. As
a result of Sobel test, depression significantly gave
mediation effect on the relation between sleep disorder and fatigue (Z=6.097, p<.000).
Conclusion
ŠŠ
These results suggest that there is a need to develop tailored program for reproductive aged women
with psoriasis to decrease fatigue considering simultaneous management of depression and sleep
disorder.
Reliability and Validity of the Persian Version of the Decisional
Conflict Scale in Selecting Mode of Child Delivery
A. Aslani1, S. Eslami2,5 , F. Tara3,
L. Ghalichi4 , A. Abu-hanna5
1_ Health Human Resources Research Center, School of Management
& Information Sciences, Shiraz University of Medical Sciences,
Shiraz, Iran
2_ Pharmaceutical Research Center, School of Pharmacy, Mashhad
University of Medical Sciences, Mashhad, Iran.
3_ Research Center for Patient Safety, Mashhad University of Medical
Sciences, Mashhad, Iran
4_ Mental Health Research Center, Iran University of Medical Sciences,
Tehran, Iran
5_ Department of Medical Informatics, Academic Medical Center,
University of Amsterdam, Amsterdam, the Netherlands
ABSTRACT
Objective
ŠŠ
Results
ŠŠ
Methods
ŠŠ
Conclusions
ŠŠ
Women’s estimates of the risks and benefits of
modes of delivery vary, and many pregnant women
unnecessarily choose for a caesarean delivery. Because some delivery options have undesirable outcomes, pregnant women may experience uncertainty and decisional conflict when facing delivery mode
choices.
A valid tool is needed to identify decisional conflicts
and their contributing factors when facing delivery
choices. The Decisional Conflict Scale (DCS) is a
16-item, self-administered questionnaire consisting
of 5 subscales developed to assess peoples’ decisional conflict. The low literacy version of the DCS
tool contains a subset of 10 items from the full DCS
in "yes, no, unsure" response format. The aim of the
present study is to investigate the reliability and validity of the Persian language version of the full and
low literacy version of the DCS.
The DCS was translated from English to Persian
and then back translated by two English language
experts. The translated version was used with 212
pregnant women in Mashhad, Iran. The scale was
statistically assessed for reliability and construct
validity including factor analysis and the Cronbach
alpha coefficient.
212 women with an average age of 26.7 (SD=7)
years, and median gestational age of 29 weeks, who
received prenatal care from public centers and private offices, were sampled. Factorial analysis identified four factors (or subscales) in pregnant women
preferring cesarean section, corresponding to: feeling informed (feeling clear about values included in
this factor); being supported; being uncertain; and
perceived effective decision making. Three subscales, in the group selecting vaginal delivery (being
uncertain; and effective decision making were in
one) were extracted. We also analysed the low literacy version of the DCS tool and found 2 subscales,
feeling informed, and being supported. The reliability
coefficients of the DCS tool were 0.87 and 0.82 in
the two versions, respectively. The Persian versions
of the DCS were not clustered in the same way as
the original version of the DCS.
The Persian version of the DCS had good construct
validity and internal consistency. This version of the
DCS was not clustered in the same way as the original DCS version. Two subscales, feeling informed
and feeling clear about values, may not reflect distinct factors in pregnant women deciding on mode
of delivery (all versions).
Keywords: decisional conflict; decision making; child delivery mode; validity; reliability
122
2016 Conference of
Asia Pacific Association for Medical Informatics
Using Simulation to assess user needs for the Point of Care (POC)
Robot development
Hyeongsuk Lee, RN, MSN1,2, Jeongeun Kim, RN, PhD, INS1,3,4, Sukwha
Kim, MD, PhD, FBSE5, Meihua Piao, RN, M.Sc3, Jisan Lee, RN, M.Sc3,
Ahjung Byun, RN3, Hyoun-Joong Kong, PhD6, Ha Min Sung, MD5
ABSTRACT
1_ College of Nursing, Seoul National University, Seoul, Republic of Korea
2_ Seoul Women’s College of Nursing, Seoul, Republic of Korea
3_ Research Institute of Nursing Science, Seoul National University, Seoul,
Republic of Korea
4_ Interdisciplinary Program of Medical Informatics, Seoul National University,
Seoul, Republic of Korea
5_ Department of Reconstructive Plastic Surgery, Seoul National University
College of Medicine, Seoul, Republic of Korea
6_ Department of Biomedical Engineering, Chungnam National University
College of Medicine and Chungnam National University Hospital, Daejeon,
Republic of Korea
Objective
ŠŠ
The adoption of telehealth technologies eliminates
location as a barrier to accessing quality treatment and care. This opens up new opportunities for
healthcare providers to engage with patients and
other professions across distances in real-time.
Point of care (POC) means healthcare providers deliver health services to patients at the time of care.
In this sense, we are currently developing the POC
robot as a tool for teleconsultation and interprofessional education. Thus, the purpose of this simulation was to assess needs of the POC robot and to
define the essential elements and functions of robot
prior to robot system development.
Methods
ŠŠ
We went through the following steps in order to derive specific user requirements and barriers in real
medical environment. Firstly, target medical fields
were selected and basic domains of the POC robot
were defined. Secondly, we used contextual inquiry
methodology that combines interviews and field observations to gain significant information concerning
the user’s context, the equipment in use, and the
place of work (Beyer, H., et al., 1997). Thirdly, we organized scenes, and mapped expected functionality
of the POC robot to each scene. Fourthly, a simulated patient was defined and medical records of the
patient were described. Lastly, clinicians reviewed
scenarios and completed final scenarios for the simulation. A total of 9 researchers (two clinicians, four
nurses, and three system developers) involved to
conduct the simulation. Three nurses and three cli-
nicians participated as observers to evaluate the robot. The simulation took place in a simulation center
with an emergency department, an operation room,
and general wards. All the simulations were filmed.
Results
ŠŠ
We selected department of neurology, dermatology
and pathology as target medical fields according
to demands of health professionals of the study
hospital. Three major functionalities tested in the
simulation: 1) image acquisition using a super wide
angle camera and a high resolution camera, 2) image transmission and display, 3) remote control of
robot arms, 4) mobility of the robot. A developed
scenario was consisted four part, and each part had
3 to 4 scenes. A simulated patient was female of
52, and she admitted an emergency department for
headache and memory impairment. Four research
assistants took roles as two clinicians, a nurse, and
a family caregiver. Observers reported that the POC
robot is highly acceptable. Also, they derived some
errors such as space conflict, communication errors,
and safety issues.
Conclusion
ŠŠ
We found that it was feasible to use the POC robot
for teleconsultation. While there are challenges,
incorporating this technology into healthcare services will enhance communication and teamwork
skills across distances, and will facilitate teamwork.
Repeated simulations will be needed to evaluate improved performances.
Keywords: Teleconsultation, Point of Care, Telepresence robot, Bedside robot
123